In order to avoid the backtracking operation of AO* algorithm and Rollout algorithm， and to overcome the problem of poor optimality of information entropy algorithm， and achieve the purpose of fault detection and isolation of equipment with low test cost， a diagnostic strategy construction method combining one-step and multi-step optimization is proposed. Based on the similarity between diagnostic strategy construction and noiseless coding， a one-step or multi-step optimal search method is selected according to the probability distribution of the fault state fuzzy set. When the fault with the highest probability in the fuzzy set of fault states meets certain conditions， it is prioritized for isolation and the corresponding test sequence is generated so that the fault with higher priori probability has less test steps and costs, so as to reduce the overall cost of diagnosis. Through the analysis and verification of examples， the concrete process of constructing the diagnostic strategy is given， and the effectiveness of the method is proved.