The invention discloses a super-large-scale distributed network computing method, which comprises the following steps that: S101, general computing provides an underlying technical protocol, and the general computing consists of S102, network communication for simulating neural form transmission, S103, distributed network computing and S104, a block chain distributed storage network; S105, minimumunit function modules required in the industry are formed according to the underlying technical protocol, wherein the minimum unit function modules form a modular protocol group; and S106, differentmodular protocol combinations are finally selected to complete different requirements of each industry field in the proprietary calculation. Compared with a traditional C / S transmission mode, a centralized computing system and centralized cloud storage, the technical scheme of adopting a peer-to-peer network computing mode, a weak centralized computing mode and full-node distributed storage provided by the invention has the advantages of high robustness, light network weight, high safety of a file system, high load of a server, low cost and the like.