| With the rapid development of the information technology, the area of software application become wider. But, the software reliability and quality are not improved correspondingly. This is a key factor to slow down the pace of the software industry developing. The induetrail process-control and traffic control, finance and medical system need the software systems with high reliability and quality. The software invalidation in these fields often leads to great economic loss. Software testing technology can estimate and improve software reliability, and can improve the software quality effectively. For the important effect in software quality assurance and software reliability validation areas, the software testing technology is payed great attention and is widely researched by researchers.SOA (Service-Oriented Architecture) is a new architecture for distributed software developing. In application system which use SOA architecture, the system function composed by a group of services with loose coupling, sites transparent, unifying defining interfaces. This technology gained widely acceptance for its important using in resolving enterprise information integration problem. The SOA and traditional software architecture have big diffience. SOA system is hard to be tested. There have none solution in this problem.The testing activities were divided into three aspect in this paper. The relation of the three aspect was analysed. For the mutation testing's success in the SOA testing, this paper put forward an improved services testing method based contract mutation. This method avoid a great defect in the equifinality class input decision for traditional mutation testing method and achieve a better effect in real experiments. In the foundation of basic contract mutation testing method, a service mutation testing method based on aritificial immunity system was given. Through aritificial immunity system, the mutants selection will become more effective and the testing efficiency will be promoted. Finally, the theory model for mutants coupling effect is discussed. An important principle for mutants optimization is given for the theory foundation of the SOA services mutation testing optimization. |