Software reliability prediction can predict the fault-prone modules at the early age of sofrware development. And it is important to improve the credibility of the software. In order to improve the effect of software reliability prediction, this paper proposes a PSOABC-SVM model to predict software reliability, and puts forward a model of predicting the software reliability based on PSOABC-SVM. The experimental results show that this model can achieve more precise prediction results than other prediction models such as BP neural network and PSO-SVM. The PSOABC-SVM model is more applicable for software reliability prediction.