Abstract: 【Objective】 Thus laying a theoretical foundation for the accumulation of SO₂ in grapes based on linear regression analysis. 【Methods】 Eleven grapes, including Flame Seedless, Summer Black, Manicure Finger, Thompson Seedless, Kyoho, Red Globe, Sweet Sapphire, Sunnny Rose, Munage, Crimson, Zizhenxiang, were studied by using correlation and multiple linear regression analysis methods to explore the relationship between stem length, width, surface area, water content, skin hole area and SO₂ accumulation in fruit stalks. 【Results】 The Pearson correlation showed that water content and skin hole area were positively correlated with SO₂ accumulation in fruit stalks. The multiple linear regression equation for predicting SO₂ accumulation in fruit stem was constructed by fumigation concentration (X₁), moisture content (X₂) and skin hole area (X₃). The cumulative prediction model Y= 0.31X₁ + 1,214.08X₂ + 1,261.34X₃-1,040.50 was obtained. The coefficient of determination (R²) of the regression equation was 0.962, The F test of significance corresponds to a P of 0, and the independent variable had a limiting influence on the dependent variable (P < 0.01). The results of regression standardized residual analysis showed that the equation conformed to normal distribution and had a high degree of fit. The average relative error between the predicted value and the actual value of the model was 0.04%, which was a low error. 【Conclusion】 The prediction structure of water content and SO₂ accumulation in fruit stem by using multiple linear regression model is more accurate, the error is lower, and it has high feasibility.