In open environment, the system behaviors are diversified and complicated, but traditional static models are mostly implemented before the deployment. They are based on certain invariance assumptions of the operating environment and user needs. We study on the monitoring of software system to find the differences of software reliability between testing phase and running phase, combining static module-based software reliability model to evaluate the system reliability during the running phase. Gives the risk factors that decrease module reliability and proposes an effective method to reallocate the resources. Through the experimental data, verifying the effectiveness of the proposed method.