Abstract: The estimation of the construction cost of urban rail transit system can control and optimize the construction cost of urban rail transit system during the design.In view of the disadvantages of traditional cost estimation model, such as large amount of calculation, miscellaneous and tedious calculation methods, etc., based on the cost data in the construction stage for multiple operating urban rail transit lines, this paper used the method of data expansion to establish the cost data set, in the case of selecting a small number of key cost indicators, established the ELM (Extreme Learning Machine) model, estimates the construction cost of urban rail transit system.The estimation results show that the MAPE(Mean Absolute Percentage Error)of the construction cost estimation model of urban rail transit system based on ELM is less than 6%.The estimated results are in good agreement with the actual data.The method adopted in this paper is scientific and effective. It can meet the engineering needs of estimating the construction cost of urban rail transit system.