diff --git a/redash/models/__init__.py b/redash/models/__init__.py
index 7cc551b659..7f0518cf4d 100644
--- a/redash/models/__init__.py
+++ b/redash/models/__init__.py
@@ -2,6 +2,7 @@
 import datetime
 import logging
 import numbers
+import re
 import time
 
 import pytz
@@ -644,6 +645,43 @@ def outdated_queries(cls):
 
         return list(outdated_queries.values())
 
+    @classmethod
+    def _do_multi_byte_search(cls, all_queries, term, limit=None):
+        # term examples:
+        #    - word
+        #    - name:word
+        #    - query:word
+        #    - "multiple words"
+        #    - name:"multiple words"
+        #    - word1 word2 word3
+        #    - word1 "multiple word" query:"select foo"
+        tokens = re.findall(r'(?:([^:\s]+):)?(?:"([^"]+)"|(\S+))', term)
+        conditions = []
+        for token in tokens:
+            key = None
+            if token[0]:
+                key = token[0]
+
+            if token[1]:
+                value = token[1]
+            else:
+                value = token[2]
+
+            pattern = f"%{value}%"
+
+            if key == "id" and value.isdigit():
+                conditions.append(cls.id.equal(int(value)))
+            elif key == "name":
+                conditions.append(cls.name.ilike(pattern))
+            elif key == "query":
+                conditions.append(cls.query_text.ilike(pattern))
+            elif key == "description":
+                conditions.append(cls.description.ilike(pattern))
+            else:
+                conditions.append(or_(cls.name.ilike(pattern), cls.description.ilike(pattern)))
+
+        return all_queries.filter(and_(*conditions)).order_by(Query.id).limit(limit)
+
     @classmethod
     def search(
         cls,
@@ -664,12 +702,7 @@ def search(
 
         if multi_byte_search:
             # Since tsvector doesn't work well with CJK languages, use `ilike` too
-            pattern = "%{}%".format(term)
-            return (
-                all_queries.filter(or_(cls.name.ilike(pattern), cls.description.ilike(pattern)))
-                .order_by(Query.id)
-                .limit(limit)
-            )
+            return cls._do_multi_byte_search(all_queries, term, limit)
 
         # sort the result using the weight as defined in the search vector column
         return all_queries.search(term, sort=True).limit(limit)
@@ -678,13 +711,7 @@ def search(
     def search_by_user(cls, term, user, limit=None, multi_byte_search=False):
         if multi_byte_search:
             # Since tsvector doesn't work well with CJK languages, use `ilike` too
-            pattern = "%{}%".format(term)
-            return (
-                cls.by_user(user)
-                .filter(or_(cls.name.ilike(pattern), cls.description.ilike(pattern)))
-                .order_by(Query.id)
-                .limit(limit)
-            )
+            return cls._do_multi_byte_search(cls.by_user(user), term, limit)
 
         return cls.by_user(user).search(term, sort=True).limit(limit)