Emergency material demand forecast of earthquake disaster includes pre-disaster and post-disaster forecast, but study on the pre-disaster forecast is relatively rare at present, this paper focus on the problem of the pre-disaster forecast using CBR (case-based reasoning) method, the purpose is to control the emergency material reserve scale at a reasonably range. First of all, using the analytic hierarchy process (AHP) to calculate the weight of each characteristic attribute for the earthquake disaster emergency demand case. Then according to the different type of characteristic attribute, calculating the similarity degree between characteristic attribute respectively. Next, using the nearest neighbor method to calculate the similarity degree between cases. Finally, using the algorithm based on the analytic method to revise the demand data of similar case set, to get reasonable demand forecast result, the demand forecast result can be used as a decisionmaking basis that control the emergency material reserve scale.