As an MFI-type zeolite, ZSM-5 zeolite has wide applications in industry, such as in the fine chemical, petrochemical, and coal chemical industries. However, shape control of ZSM-5 nanocrystals constitutes one of the major challenges of current nanotechnology. Here, the MFI framework structure was used as the theoretical model of pure silicon ZSM-5 to investigate the surface energy and Wulff shape. The models with different crystal surfaces were simulated by molecular dynamics (MD) with the assistance of machine learning potentials (MLPs). The factors influencing the crystal surface energy, such as temperature, pH, and ionic concentration, have been studied in detail. Depending on the calculated surface energies, the crystal surface morphology and its ratio were obtained by means of the Wulff theorem. The results show that the area in the equilibrium shape of the (110) surface is usually the largest, and its proportion varies with external conditions. A high temperature and high concentration of the aluminum source promoted the growth of the (110) crystal surface, and the theoretical value of the crystal surface ratio was as high as 90%. This study provides theoretical insight into the synthesis of zeolites with different morphologies of all-silicon or low-aluminum.