diff --git a/jvm-packages/xgboost4j/src/test/java/ml/dmlc/xgboost4j/java/BoosterImplTest.java b/jvm-packages/xgboost4j/src/test/java/ml/dmlc/xgboost4j/java/BoosterImplTest.java
index e6fa7d709cff..41a355939bcc 100644
--- a/jvm-packages/xgboost4j/src/test/java/ml/dmlc/xgboost4j/java/BoosterImplTest.java
+++ b/jvm-packages/xgboost4j/src/test/java/ml/dmlc/xgboost4j/java/BoosterImplTest.java
@@ -185,6 +185,38 @@ public void testDescendMetricsWithBoundaryCondition() {
     }
   }
 
+  @Test
+  public void testEarlyStoppingForMultipleMetrics() {
+    Map<String, Object> paramMap = new HashMap<String, Object>() {
+      {
+        put("max_depth", 3);
+        put("silent", 1);
+        put("objective", "binary:logistic");
+        put("maximize_evaluation_metrics", "true");
+      }
+    };
+    float[][] metrics = new float[3][5];
+    for (int i = 0; i < 3; i++) {
+      for (int j = 0; j < 5; j++) {
+        metrics[0][j] = j;
+      }
+    }
+    boolean onTrack = XGBoost.judgeIfTrainingOnTrack(paramMap, 5, metrics, 4);
+    TestCase.assertTrue(onTrack);
+    for (int i = 0; i < 5; i++) {
+      metrics[0][i] = 5 - i;
+    }
+    // when we have multiple datasets, the training metrics is not considered
+    onTrack = XGBoost.judgeIfTrainingOnTrack(paramMap, 5, metrics, 4);
+    TestCase.assertTrue(onTrack);
+    for (int i = 0; i < 5; i++) {
+      metrics[1][i] = 5 - i;
+    }
+    // if any metrics off, we need to stop
+    onTrack = XGBoost.judgeIfTrainingOnTrack(paramMap, 5, metrics, 4);
+    TestCase.assertFalse(onTrack);
+  }
+
   @Test
   public void testDescendMetrics() {
     Map<String, Object> paramMap = new HashMap<String, Object>() {