The gravitational lensing effect is one of the most effective probes for detecting the distribution of mass in galaxy clusters. To date, the methods of gravitational lensing mass reconstructions can be divided into two categories, parametric methods and non-parametric methods. However, due to the model assumptions and computational resolution, the available reconstruction methods still have a number of problems that need to be solved. The method of gravitational lens mass reconstruction based on Shapelets basis function is used to reconstruct the gravitational lens mass. The Shapelets basis function is used to decompose the gravitational lens potential, and the basis function coefficients are iteratively calculated by the constraints of the multiple image position and the background galaxy ellipticity distortion of gravitational lensing. The new method is validated by fitting the mock data from NFW (Navarro, Frenk and White) lensing model. The results show the gravitational lensing mass distribution can be reconstructed reasonably by our new method and the source position can be calculated accurately. Our new method provides a flexible and efficient method for gravitational lensing mass reconstruction.