diff --git a/rdagent/components/coder/CoSTEER/knowledge_management.py b/rdagent/components/coder/CoSTEER/knowledge_management.py
index 820c21381f..c486e5aa69 100644
--- a/rdagent/components/coder/CoSTEER/knowledge_management.py
+++ b/rdagent/components/coder/CoSTEER/knowledge_management.py
@@ -519,13 +519,10 @@ def analyze_error(
         else:
             error_list = []
             for error_content in error_contents:
-                for error_node in all_error_nodes:
-                    if error_content == error_node.content:
-                        error_list.append(error_node)
-                    else:
-                        error_list.append(error_content)
-                    if error_list[-1] in error_list[:-1]:
-                        error_list.pop()
+                matched_node = self.knowledgebase.graph.find_node(content=error_content, label="error")
+                error_item = error_content if matched_node is None else matched_node
+                if error_item not in error_list:
+                    error_list.append(error_item)
 
             return error_list
 
diff --git a/test/utils/coder/test_costeer_analyze_error.py b/test/utils/coder/test_costeer_analyze_error.py
new file mode 100644
index 0000000000..c50a096ed3
--- /dev/null
+++ b/test/utils/coder/test_costeer_analyze_error.py
@@ -0,0 +1,60 @@
+from types import SimpleNamespace
+
+import pytest
+
+from rdagent.components.coder.CoSTEER.knowledge_management import (
+    CoSTEERRAGStrategyV2,
+)
+from rdagent.components.knowledge_management.graph import (
+    UndirectedGraph,
+    UndirectedNode,
+)
+
+ROWS_ERROR = "The source dataframe and the ground truth dataframe have different rows count."
+TOLERANCE_ERROR = "Some values differ by more than the tolerance of 1e-6."
+EXECUTION_FEEDBACK = 'File "factor.py", line 3, in <module>\n    x = 1 / 0\nZeroDivisionError: division by zero'
+EXECUTION_ERROR = "ErrorType: ZeroDivisionError\nError line: x = 1 / 0"
+
+
+def _error(content: str) -> UndirectedNode:
+    return UndirectedNode(content=content, label="error")
+
+
+def _strategy(*nodes: UndirectedNode) -> CoSTEERRAGStrategyV2:
+    graph = UndirectedGraph()
+    graph.nodes = {node.id: node for node in nodes}
+    strategy = CoSTEERRAGStrategyV2.__new__(CoSTEERRAGStrategyV2)
+    strategy.knowledgebase = SimpleNamespace(graph=graph)
+    return strategy
+
+
+@pytest.mark.offline
+@pytest.mark.parametrize(
+    ("feedback", "feedback_type", "content"),
+    [
+        (ROWS_ERROR, "value", ROWS_ERROR),
+        (EXECUTION_FEEDBACK, "execution", EXECUTION_ERROR),
+        ("Execution timed out after 600 seconds.", "execution", "Undefined Error"),
+    ],
+    ids=["value", "execution", "undefined"],
+)
+def test_analyze_error_returns_matched_node_once(feedback: str, feedback_type: str, content: str) -> None:
+    matched = _error(content)
+    strategy = _strategy(_error("A different previous error."), matched)
+    assert strategy.analyze_error(feedback, feedback_type=feedback_type) == [matched]
+
+
+@pytest.mark.offline
+@pytest.mark.parametrize("graph_order", ["parsed", "reversed"])
+def test_analyze_error_orders_matched_nodes_by_feedback(graph_order: str) -> None:
+    rows, tolerance = _error(ROWS_ERROR), _error(TOLERANCE_ERROR)
+    strategy = _strategy(rows, tolerance) if graph_order == "parsed" else _strategy(tolerance, rows)
+    feedback = f"{ROWS_ERROR}\n{TOLERANCE_ERROR}\n{ROWS_ERROR}"
+    assert strategy.analyze_error(feedback, feedback_type="value") == [rows, tolerance]
+
+
+@pytest.mark.offline
+def test_analyze_error_keeps_unmatched_error_in_order() -> None:
+    rows = _error(ROWS_ERROR)
+    strategy = _strategy(_error("A different previous error."), rows)
+    assert strategy.analyze_error(f"{TOLERANCE_ERROR}\n{ROWS_ERROR}", feedback_type="value") == [TOLERANCE_ERROR, rows]