diff --git a/data_diff/sqeleton/databases/mysql.py b/data_diff/sqeleton/databases/mysql.py
index fd4bc295..a901c0c6 100644
--- a/data_diff/sqeleton/databases/mysql.py
+++ b/data_diff/sqeleton/databases/mysql.py
@@ -1,3 +1,4 @@
+import logging
 from ..abcs.database_types import (
     Datetime,
     Timestamp,
@@ -21,6 +22,8 @@
 from .base import MD5_HEXDIGITS, CHECKSUM_HEXDIGITS, TIMESTAMP_PRECISION_POS, Mixin_Schema, Mixin_RandomSample
 from ..queries.ast_classes import BinBoolOp
 
+logger = logging.getLogger("mysql")
+
 
 @import_helper("mysql")
 def import_mysql():
@@ -127,6 +130,7 @@ def __init__(self, *, thread_count, **kw):
         self._args = kw
 
         super().__init__(thread_count=thread_count)
+        self.check_charset()
 
         # In MySQL schema and database are synonymous
         try:
@@ -134,10 +138,23 @@ def __init__(self, *, thread_count, **kw):
         except KeyError:
             raise ValueError("MySQL URL must specify a database")
 
+    def check_charset(self) -> None:
+        if logging.getLogger().level == logging.DEBUG:
+            try:
+                char_set_result = self.query("SELECT @@character_set_client;")
+                logger.debug(f"charset: {char_set_result.rows}")
+                collation_result = self.query("SELECT @@collation_connection;")
+                logger.debug(f"collation: {collation_result.rows}")
+            except Exception as ex:
+                logger.warning(f"Failed to check the charset: {ex}")
+
     def create_connection(self):
         mysql = import_mysql()
         try:
-            return mysql.connect(charset="utf8", use_unicode=True, **self._args)
+            conn = mysql.connect(charset="utf8mb4", use_unicode=True, **self._args)
+            conn.set_charset_collation(charset="utf8mb4", collation="utf8mb4_0900_ai_ci")
+            return conn
+
         except mysql.Error as e:
             if e.errno == mysql.errorcode.ER_ACCESS_DENIED_ERROR:
                 raise ConnectError("Bad user name or password") from e