Abstract: The software reliability model is one of the best approaches to predict. This purer analyzed and evaluated software reliability quantitatively. It is very important to infer software reliability by reasonably modeling and incorporating software failure data as well as other prior information. The basic concepts of software reliability models were presented. More than thirty different models proposed recently were analysed. Their predictive validity and applicability were discussed in detail. Finally,future research directions and potential applications of software reliability models were pointed out.

Key words: Software reliability models, Stochastic process, Unascertained theory, Machine learning, Bayesian inference,Chaos