In order to implement reliable peer-to-peer（P2P） computation in dynamic and unstable environments,a reliable P2P computing model based on the probabilistic redundant scheduling is established.In this model,the relationship among the success rate of P2P scheduling,the resource stability and the resource redundancy is used to obtain an optimal redundancy of task scheduling,improve the scheduling performance,and guarantee the reliability of P2P scheduling by probability.Moreover,a reliable probabilistic redundancy-based algorithm for task scheduling is designed,and the effectiveness of the proposed model and algorithm are validated by experiments.