Bridge influence line is an important tool to study the bridge response under the moving loads, which contains tremendous structural information. The structural deterioration during the service life would induce the variation of the bridge influence line, which make it essential to identify the influence line exactly from the field measurement data under moving vehicle loads. This paper establishes the underdetermined influence line identification model for bridge under multi-axle vehicle excitation, and the bridge influence line is identified based on regularized least square QR decomposition method. Then an example of a four spans girder bridge is illustrated to prove the validity of the method, the accuracy and computational complexity of different methods in this condition is compared. Results of this study provide an effective identification method of bridge influence line with very high computational efficiency and identification accuracy.