Scientific, reasonable and accurate prediction of earthquake emergency supplies demand can provide a basis for rapid decision-making for the collection and scheduling of emergency supplies, so as to improve the efficiency of earthquake relief work. Based on the destructive earthquake disaster data in China from 2008 to 2020, we utilize the regression analysis to study the demand prediction of typical emergency supplies related to the affected population by the earthquake.The quartile method of the difference measure results is adopt to adjust the prediction results of the number of disaster victims through hypothesis inspection and model modification.The last, the quantity demanded of drinking water and clothes for emergency supplies are calculated. The experimental results show that the prediction model is instructive to reaching a rational and practicable level.