Distributed systems and applications require efficient and effective techniques (e.g. self (re)configuration, self-healing, etc.) for ensuring safety, security and more generally dependability properties, as well as convergence. The complexity of these systems is increased by features like dynamic (changing) topology, interconnection of heterogeneous components or failures detection. This paper presents a methodology for verifying protocols and satisfying safety and convergence requirements of the distributed self- systems. The self- systems are based on the idea of managing complex infrastructures, software, and distributed systems, with or without minimal user interactions. Correct-by-construction and service-as-event paradigms are used for formalizing the system requirements, where the formalization process is based on incremental refinement in EVENT B. Moreover, this paper describes a fully mechanized proof of correctness of the self-* systems along with an interesting case study related to the P2P-based self healing protocol.