The scheme of condition detection and fault diagnosis system based on cloud platform was studied,and the strategy of dual software architecture based on NETWORX and DJANGO was proposed to solve the problem of network integration of monitoring evaluation and fault diagnosis,and to realize the purpose of monitoring multiple devices.The field control system with PLC as the core was designed; the NETWORX architecture could easily exchange data with various internet of things acquisition systems,so it was used to realize the remote monitoring program of cloud platform; the programs of state detection and fault diagnosis were designed by using DJANGO of Python.The results show that the acquisition accuracy of the proposed system characteristic parameters is within 1%,and the control and display monitoring functions both meet the design requirements.The intelligent algorithms of classification and regression trees（CART）,SVM and MLP were compared with the experience fault data.The results show that the fault correct rate of CART algorithm,SVM algorithm and MLP algorithm is 91.3%,73.2% and 86.2% respectively,which proves that the state monitoring and fault diagnosis system based on cloud platform can meet the design needs.