SQLite is an important database engine on Android platform. It has features of zero-configuration, supporting transactions and excellent portability. And it is responsible for multiple formats of data storage. With the expansion and complexity of the user data, there is higher demand upon the SQLite performance which may affect the performance of the application program and the user experience. This paper analyzes the transaction mechanism and the index structure of SQLite. By means of turning on transaction manually and creating B-tree index properly, it greatly reduces the time needed to insert and query data from Android SQLite, which improves SQLite performance ultimately.