Abstract: In order to improve the calculation accuracy of distribution network theoretical line loss,an intelligent calculation method based on the improved quantum genetic algorithm-core vector machine（QGA-CVM） was proposed.With the QGA-CVM,the theoretical line loss calculation is abstracted into a regression analysis,which can reach the solution;then the lines,whose theoretical losses have already been known,were structured into a sample set as the data source of CVM.In order to obtain the fitting function of the regression analysis,the sample set was trained by CVM.In this training process,the optimal training parameters can be searched by QGA to reduce the blind training parameters selection of CVM,and improve the line loss calculation accuracy based on the QGA-CVM.Finally,the experiments proved the effectiveness of the proposed theoretical line loss calculation method.Compared with the conventional methods,the QGA-CVM method has better performance in both calculation accuracy and computing speed.