Abstract. Distributed systems and applications are becoming increasingly complex, due to factors such as dynamic topology, heterogeneity of components, failure detection. Therefore, they require effective techniques for guaranteeing safety, security and convergence. The self- systems are based on the idea of managing efficiently complex systems and architectures without user interaction. This paper presents a methodology for verifying distributed systems and ensuring safety and convergence requirements: Correct-by-construction and service-as-event paradigms are used for formalizing the system requirements using incremental refinement in E VENT B. Moreover, this paper describes a mechanized proof of correctness of the self- systems along with a case study related to the P2P-based self-healing protocol.