After examining the specific problems of testing and quality assurance for web-based applications, we propose a strategy by integrating existing testing techniques and reliability analyses in a hierarchical framework. This strategy combines various usage models for statistical testing to perform high level testing and to guide selective testing of critical and frequently used subparts or components using traditional coverage-based structural testing. Reliability analysis and risk identification form an integral part of this strategy to help assure and improve the overall reliability for web-based applications. Some preliminary results are included to demonstrate the general viability and effectiveness of our approach.