Abstract In order to improve the forecast accuracy of cigarette market demand. By establishing ARIMA, VAR, SVR and the cigarette demand model under habitual consumption, the demand for cigarettes in Tongchuan City is modeled and predicted. Based on the idea of integrated method, the prediction results of the above four models are integrated and analyzed. Results show that the absolute relative error of the integrated model in forecasting cigarette sales in 2017-2018 is controlled within 1%, and the absolute value of the average relative error is controlled at about 0.8%. The prediction result is significantly better than the single model.