【Objective】Quantitative analysis of temporal and spatial characteristics of agricultural non-point source pollution in Shanxi Province during 2011—2022 and prediction of its development trend. It can provides scientific reference for the prevention and control of agricultural non-point source pollution in Shanxi Province. 【Methods】The temporal and spatial characteristics of agricultural non-point source pollution in Shanxi Province were analyzed by pollution emission coefficient method and equivalent pollution load method, and ARIMA model was established to predict its development trend. 【Result】(1) In 2022, the agricultural non-point source pollution in Shanxi Province produced 4.74 ×105 t of COD, 5.82×103 t of HN3-N, 2.86×104 t of TN, and 4.18×103 t of TP. The high emission areas are in Lvliang and Yuncheng, and the low emission areas are in Taiyuan and Yangquan. The equivalent pollution load was 3.95×1010 m3. The primary pollutant is TN, and the primary pollution account is the animal husbandry account. (2) From 2011 to 2022, the development trend of the four types of pollutants showed consistency. All showed fluctuation rise, and the highest value of each pollutant appeared in 2022. The ranking of pollutant in each region is stable, Yuncheng and Lvliang have always occupied an important position. TN has been the primary pollutant for 12 years, and the animal husbandry account has been the primary pollution account, and the proportion is still rising. (3) By determining model parameters, ARIMA (1,1,2) model is established for forecast. During the forecast period, agricultural non-point source pollution in Shanxi Province will decrease slightly, and then the development trend showed a steady rise. 【Conclusion】It is necessary to further strengthen the prevention and control of key agricultural pollution accounts and key pollutants, reduce the discharge of agricultural non-point source pollution, and mitigate its rising trend.