In order to study the effects of different tillage practices on the yield of runoff and sediment in purple soil area in peanut podding stage. Field rainfall simulation experiments were carried out to study the yield of runoff and sediment in four tillage practices and one control group (longitudinal ridge, cross ridge, bunch planting, flat planting and bare land on flat slope). The sliding-rank-sum-test was used to find out the catastrophe point of runoff and sediment production process. The Mann-Kendall trend test method was used to analyze the trend of runoff and sediment yield process after the catastrophe point. The results showed that the runoff and sediment yield process was in a relatively stable state after the catastrophe point. The stable rate of runoff and sediment yield in bare land was the largest, followed by longitudinal ridge, then flat planting, bunch planting, and cross ridge was the smallest. However, the initial runoff yield time of each tillage practice was opposite to the stable rate of runoff and sediment yield. Compared to bare land on flat slope, all tillage practices showed the effects of runoff and sediment reduction. The specific performance was cross ridge > bunch planting > flat planting > longitudinal ridge.