In view of the ill-posed problems on solving moving force identification system matrix , a truncated generalized singular value decomposition(TGSVD) algorithm for moving force identification has been developed by introducing regularization matrix. The numerical simulation of two-axle moving force loads show that the proposed method TGSVD has many good properties compared with time domain method(TDM), such as better identification accuracy, robust noise immunity and less effect of response type and response combination, which is beneficial to the application of TGSVD in the field identification of dynamic axle loads on bridge.