It is necessary to estimate the operating cost for constructing an urban railway system. The present study was thus carried out to develop an estimation function of the operating cost for a MRT(Medium-sized Rapid Transit) system. We selected seven independent variables that could influence the operating cost: the rolling stocks, the number of trains in operation, the length of the lines, the number of stations, the number of passengers per day, the frequency of train operation, and the number of depots. We performed a multiple regression using Excel. Three types of regression functions were thereupon developed. All of them proved to be appropriate after comparing the results of the estimated cost. It will therefore be possible to use one of these three types, depending on the assumptions of the independent variables.