Compared with the traditional sales forecast based on human experience calculation and intuition prediction, combining existing sales data and model predictions can often accurately determine the market trend and reduce the probability of judgment error. A high degree of precision in market forecasting is essential for business leaders to develop policies and plans, increase store turnover and reduce operating costs. Based on the analysis of the historical sales data of Chinas liquor retail industry in the post-epidemic era, this study compared two existing forecasting models for forecast accuracy to determine the most suitable model for long-term sales forecasting. In this study, the sales data were collected from Suhe Bar Chain under the company of Alliance Art Group. The raw data set was screened and cleaned to suit each model, and then applied to Linear Regression Model (LRM) and Autoregressive Integrated Moving Average Model (ARIMA) to generate new forecast data. The predicted data were compared with the real sales data, and Root Mean Square Error (RMSE) was used to judge the accuracy of the model prediction. Finally, ARMIA is the better model to predict Chinas liquor retail in the post-epidemic era. According to the model predictions, reducing inventory, maintaining efficient cash flow, improving the turnover efficiency of goods, and strengthening the ability to adjust market strategy are the more suitable strategies for current liquor sales enterprises.