The invention discloses an intelligent selection method of a multi-stage data mining algorithm for high concurrency, comprising the following steps: step 1, data characteristic analysis: analyzing themass data of a cold, heat and electric hybrid energy system; 2, data preprocessing: adopting a multi-stage data process method to effectively preprocess that original data; 3, preliminary design of afeasible data mining model: selecting and designing a potentially effective mining model; 4, adjustment of the feasibility data mining model: deleting the untenable model to obtain a feasible data mining model set; 5, evaluation of the mining model: synthetically evaluating, and selecting a mining model that is most suitable for application in a target domain; 6, model deployment: putting the model into application. If the mining model cannot meet the business requirements, the mining data are adjusted and remodeling is performed. The invention realizes automatic selection and effective establishment of the data mining model, thereby improving the quality and efficiency of the model design.