Abstract：The Mw7.8 Nepal earthquake occurred on April 25, 2015, with several strong aftershocks that caused effects in the Tibetan areas of China. In order to predict the strong aftershocks of Nepal earthquake, an improved “waiting time” model using weighted least square fitting is proposed based on the theoretic residual analysis. The new model is then applied to the Nepal earthquake sequences to better predict outcomes. Results indicate that the proposed method effectively reduces model prediction errors as well as the confidence interval. Also the prediction errors are zoomed by ln10·Δt (Δt is the waiting time of the last aftershock) from model axis to real time axis. For this reason, the aftershock samples should be chosen within 3 days of the main earthquake.