To investigate the application effect of stochastic kinetic energy backscatter (SKEB) perturbation scheme, a regional ensemble forecast system (REFS) is constructed based on the WRF model. Sensitivity experiments on perturbation amplitude are conducted to find the functional characteristic of SKEB. Additionally, a synthetic model perturbation scheme which combines multiphysics (namely MPHY) and SKEB, called SKEBMPHY, is developed, and the forecast effects of the three schemes of SKEB, MPHY and SKEBMPHY are compared and evaluated. The results show that the SKEB scheme can estimate the stream function and temperature perturbation through a constant kinetic dissipation rate, and the exhibited difference of model integration is correlated to the amplitude of the kinetic dissipation rate. Perturbing the vertica1 structure is not helpful to the difference growth. The comparison results of SKEB, MPHY and SKEBMPHY schemes indicate that the SKEB has advantage on upperair wind spread growth compared to MPHY while MPHY has larger spread for lowlevel temperature than SKEB. The SKEBMPHY scheme exhibits the best perturbation growth characteristic of the three schemes not only for upperair variables but also for lowlevel variables. In addition, the SKEBMPHY scheme has the most skillful performance of the three schemes in terms of ensemble verifications. Furthermore, compared with single MPHY scheme, applying the SKEB scheme will improve precipitation forecast skill of heavy rain. All the results from this study indicate that the SKEB scheme and the SKEBbased synthetic scheme are promising.