diff --git a/frontend/nginx-locations.inc b/frontend/nginx-locations.inc
index 06d8912..780f5c5 100644
--- a/frontend/nginx-locations.inc
+++ b/frontend/nginx-locations.inc
@@ -23,6 +23,21 @@
         chunked_transfer_encoding   on;
     }
 
+    # ── LLM / AI 검증 (Ta4j 재평가 + LLM 최대 120s — 일반 /api/ 60s 초과) ───
+    location ~ ^/api/strategy/evaluation/(ai-verify|llm-test)$ {
+        proxy_pass          http://backend:8080;
+        proxy_http_version  1.1;
+        proxy_set_header    Host              $host;
+        proxy_set_header    X-Real-IP         $remote_addr;
+        proxy_set_header    X-Forwarded-For   $proxy_add_x_forwarded_for;
+        proxy_set_header    X-Forwarded-Proto $scheme;
+        proxy_connect_timeout 30s;
+        proxy_read_timeout    300s;
+        proxy_send_timeout    300s;
+        proxy_buffering       off;
+        client_max_body_size  16m;
+    }
+
     # ── GoldenChart Spring Boot Backend 프록시 ────────────────────────
     location /api/ {
         proxy_pass          http://backend:8080;
diff --git a/frontend/src/components/strategyEvaluation/StrategyEvaluationAiVerifyModal.tsx b/frontend/src/components/strategyEvaluation/StrategyEvaluationAiVerifyModal.tsx
index d71b9f2..f7116d3 100644
--- a/frontend/src/components/strategyEvaluation/StrategyEvaluationAiVerifyModal.tsx
+++ b/frontend/src/components/strategyEvaluation/StrategyEvaluationAiVerifyModal.tsx
@@ -299,7 +299,7 @@ const StrategyEvaluationAiVerifyModal: React.FC<Props> = ({
             <div className="seval-ai-verify-loading" aria-live="polite">
               <span className="btd-run-spinner seval-ai-verify-spinner" aria-hidden />
               <p>LLM이 평가 결과와 차트 시그널을 분석 중입니다…</p>
-              <p className="seval-ai-verify-loading-hint">최대 2분 정도 걸릴 수 있습니다.</p>
+              <p className="seval-ai-verify-loading-hint">Ta4j 재평가와 LLM 분석에 최대 2~3분 걸릴 수 있습니다.</p>
             </div>
           )}
 
diff --git a/frontend/src/utils/backendApi.ts b/frontend/src/utils/backendApi.ts
index 9af9e6f..015dd22 100644
--- a/frontend/src/utils/backendApi.ts
+++ b/frontend/src/utils/backendApi.ts
@@ -118,17 +118,69 @@ async function requestOrThrow<T>(path: string, init?: RequestInit): Promise<T> {
   const text = await res.text().catch(() => '');
   if (!res.ok) {
     let msg = `요청 실패 (${res.status})`;
-    try {
-      const j = JSON.parse(text) as { message?: string };
-      if (j.message) msg = j.message;
-    } catch { /* plain text */ }
-    if (!text.startsWith('{') && text) msg = text.slice(0, 200);
+    if (res.status === 504) {
+      msg = '게이트웨이 시간 초과(504). LLM 분석은 2~3분 걸릴 수 있습니다. 잠시 후 다시 시도해 주세요.';
+    } else {
+      try {
+        const j = JSON.parse(text) as { message?: string };
+        if (j.message) msg = j.message;
+      } catch { /* plain text */ }
+      if (!text.startsWith('{') && text) {
+        const plain = text.replace(/<[^>]+>/g, ' ').replace(/\s+/g, ' ').trim();
+        if (plain) msg = plain.slice(0, 200);
+      }
+    }
     throw new Error(msg);
   }
   if (res.status === 204 || !text) return undefined as T;
   return JSON.parse(text) as T;
 }
 
+/** LLM·AI 검증 등 장시간 API */
+async function requestOrThrowWithTimeout<T>(
+  path: string,
+  init?: RequestInit,
+  timeoutMs = 200_000,
+): Promise<T> {
+  const controller = new AbortController();
+  const timer = setTimeout(() => controller.abort(), timeoutMs);
+  try {
+    const res = await fetch(`${API_BASE}${path}`, {
+      ...init,
+      signal: controller.signal,
+      headers: { ...authHeaders(), ...(init?.headers ?? {}) },
+    });
+    const text = await res.text().catch(() => '');
+    if (!res.ok) {
+      let msg = `요청 실패 (${res.status})`;
+      if (res.status === 504) {
+        msg = '게이트웨이 시간 초과(504). LLM 분석은 2~3분 걸릴 수 있습니다. 잠시 후 다시 시도해 주세요.';
+      } else {
+        try {
+          const j = JSON.parse(text) as { message?: string };
+          if (j.message) msg = j.message;
+        } catch { /* plain text */ }
+        if (!text.startsWith('{') && text) {
+          const plain = text.replace(/<[^>]+>/g, ' ').replace(/\s+/g, ' ').trim();
+          if (plain) msg = plain.slice(0, 200);
+        }
+      }
+      throw new Error(msg);
+    }
+    if (res.status === 204 || !text) return undefined as T;
+    return JSON.parse(text) as T;
+  } catch (e) {
+    if (e instanceof DOMException && e.name === 'AbortError') {
+      throw new Error(
+        `요청 시간 초과(${Math.round(timeoutMs / 1000)}초). LLM이 느리면 설정 > LLM에서 타임아웃을 늘리거나 잠시 후 다시 시도하세요.`,
+      );
+    }
+    throw e;
+  } finally {
+    clearTimeout(timer);
+  }
+}
+
 async function request<T>(path: string, init?: RequestInit): Promise<T | null> {
   try {
     const res = await fetch(`${API_BASE}${path}`, {
@@ -1690,11 +1742,15 @@ export interface StrategyEvaluationAiVerifyResponse {
 export async function fetchStrategyEvaluationAiVerify(
   context: Record<string, unknown>,
 ): Promise<StrategyEvaluationAiVerifyResponse> {
-  return requestOrThrow<StrategyEvaluationAiVerifyResponse>('/strategy/evaluation/ai-verify', {
-    method: 'POST',
-    cache: 'no-store',
-    body: JSON.stringify({ context }),
-  });
+  return requestOrThrowWithTimeout<StrategyEvaluationAiVerifyResponse>(
+    '/strategy/evaluation/ai-verify',
+    {
+      method: 'POST',
+      cache: 'no-store',
+      body: JSON.stringify({ context }),
+    },
+    200_000,
+  );
 }
 
 export interface LlmConnectionTestResponse {
@@ -1706,10 +1762,14 @@ export interface LlmConnectionTestResponse {
 
 /** 설정 화면 — LLM 서버 연결 테스트 (저장된 설정 기준) */
 export async function fetchLlmConnectionTest(): Promise<LlmConnectionTestResponse> {
-  return requestOrThrow<LlmConnectionTestResponse>('/strategy/evaluation/llm-test', {
-    method: 'POST',
-    cache: 'no-store',
-  });
+  return requestOrThrowWithTimeout<LlmConnectionTestResponse>(
+    '/strategy/evaluation/llm-test',
+    {
+      method: 'POST',
+      cache: 'no-store',
+    },
+    130_000,
+  );
 }
 
 /** 전략 DSL에 포함된 평가 시간봉 목록 (실시간 체크·STOMP 구독용) */