Background:　With　the　rapid　development　of　computer　and　Internet　technologies,　the　measurement　of　software　reliability　has　become　a　key　point.　Software　reliability　modeling　is　an　important　way　to　measure　software　reliability.　However,　there　are　many　disadvantages　in　the　classic　model.　Therefore,　this　paper　reviews　current　researches　in　software　reliability　and　discusses　ways　to　improve　the　model　from　a　new　perspective.　Objective:　This　paper　reviews　the　current　researches　on　software　reliability,　analyzes　the　existing　patented　technology,　and　points　out　some　future　directions　to　improve　the　current　models　to　provide　a　more　accurate　and　efficient　way　for　software　reliability　measurement.　Methods:　This　study　introduces　the　general　methods　of　software　reliability　evaluation,　analyzes　the　disadvantages　of　these　methods　in　practical　application,　draws　lessons　from　the　patent,　and　puts　forward　some　new　ideas　to　solve　the　problems.　Results:　Disadvantages　of　current　researches　are　identified　based　on　literature　review.　In　addition,　two　suggestions　are　given　to　improve　the　current　models.　Conclusion:　Although　the　software　reliability　has　made　great　progress,　there　are　still　many　things　to　be　studied　in　the　future.　The　problems　of　software　reliability　in　real　life　should　be　gradually　solved.　©　2021　Bentham　Science　Publishers.