diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt
index aec321005770..ef1e808c2559 100644
--- a/lucene/CHANGES.txt
+++ b/lucene/CHANGES.txt
@@ -182,6 +182,8 @@ Optimizations
 
 * GITHUB#14863: Perform scoring for 4, 7, 8 bit quantized vectors off-heap. (Kaival Parikh)
 
+* GITHUB#15261: Implement longValues for MultiFieldNormValues to speedup CombinedQuery (Ge Song)
+
 Bug Fixes
 ---------------------
 * GITHUB#14161: PointInSetQuery's constructor now throws IllegalArgumentException
diff --git a/lucene/core/src/java/org/apache/lucene/search/MultiNormsLeafSimScorer.java b/lucene/core/src/java/org/apache/lucene/search/MultiNormsLeafSimScorer.java
index 65609921bc93..179872e52435 100644
--- a/lucene/core/src/java/org/apache/lucene/search/MultiNormsLeafSimScorer.java
+++ b/lucene/core/src/java/org/apache/lucene/search/MultiNormsLeafSimScorer.java
@@ -18,6 +18,7 @@
 
 import java.io.IOException;
 import java.util.ArrayList;
+import java.util.Arrays;
 import java.util.Collection;
 import java.util.HashSet;
 import java.util.List;
@@ -25,6 +26,7 @@
 import java.util.Set;
 import org.apache.lucene.index.LeafReader;
 import org.apache.lucene.index.NumericDocValues;
+import org.apache.lucene.internal.hppc.FloatArrayList;
 import org.apache.lucene.search.CombinedFieldQuery.FieldAndWeight;
 import org.apache.lucene.search.similarities.Similarity.BulkSimScorer;
 import org.apache.lucene.search.similarities.Similarity.SimScorer;
@@ -127,8 +129,10 @@ public float score(int doc, float freq) throws IOException {
    */
   public void scoreRange(DocAndFloatFeatureBuffer buffer) throws IOException {
     normValues = ArrayUtil.growNoCopy(normValues, buffer.size);
-    for (int i = 0; i < buffer.size; i++) {
-      normValues[i] = getNormValue(buffer.docs[i]);
+    if (norms != null) {
+      norms.longValues(buffer.size, buffer.docs, normValues, 1L);
+    } else {
+      Arrays.fill(normValues, 0, buffer.size, 1L);
     }
     bulkScorer.score(buffer.size, buffer.features, normValues, buffer.features);
   }
@@ -145,6 +149,7 @@ public Explanation explain(int doc, Explanation freqExpl) throws IOException {
 
   private static class MultiFieldNormValues extends NumericDocValues {
     private final NumericDocValues[] normsArr;
+    private float[] accBuf = FloatArrayList.EMPTY_ARRAY;
     private final float[] weightArr;
     private long current;
     private int docID = -1;
@@ -193,5 +198,33 @@ public int advance(int target) {
     public long cost() {
       throw new UnsupportedOperationException();
     }
+
+    @Override
+    public void longValues(int size, int[] docs, long[] values, long defaultValue)
+        throws IOException {
+      if (accBuf.length < size) {
+        accBuf = new float[ArrayUtil.oversize(size, Float.BYTES)];
+      } else {
+        Arrays.fill(accBuf, 0f);
+      }
+
+      for (int i = 0; i < normsArr.length; i++) {
+        // this code relies on the assumption that document length can never be equal to 0,
+        // so we can use 0L to indicate whether we have a norm value or not
+        normsArr[i].longValues(size, docs, values, 0L);
+        float weight = weightArr[i];
+        for (int j = 0; j < size; j++) {
+          accBuf[j] += weight * LENGTH_TABLE[Byte.toUnsignedInt((byte) values[j])];
+        }
+      }
+
+      for (int i = 0; i < size; i++) {
+        if (accBuf[i] == 0f) {
+          values[i] = defaultValue;
+        } else {
+          values[i] = SmallFloat.intToByte4(Math.round(accBuf[i]));
+        }
+      }
+    }
   }
 }