Abstract: The quantitative estimation of recharge by precipitation infiltration plays an important role in the evaluation of regional water resources. In arid and semi-arid areas, the annual amount of recharge by precipitation infiltration usually depends on the characteristics of several concentrated precipitation events. Therefore, research on the mechanism of sub-precipitation infiltration recharge (sub-recharge) and the establishment of an estimation method of the amount of sub-recharge can improve the scientific and accurate evaluation of annual infiltration recharge in arid and semi-arid areas. This study investigated the sub-recharge mechanism using the piedmont region of the North China Plain as an example. Large amounts of soil water content, soil matric potential, recharge flux to groundwater, and precipitation of the Luancheng Agro-Ecosystem Experimental Station, Chinese Academy of Sciences (LC Station), and Ranzhuang Water Resources Experiment Station, Hebei Province (RZ Station) were used. Firstly, the depth of potential recharge occurrence, that is, the zero-flux plane based on the change characteristics of the soil water potential gradient, was determined; and the patterns of precipitation infiltration recharge were analyzed. The depths of potential recharge were determined to be 2 m at the LC Station and 3 m at the RZ Station. The difference in the potential recharge occurrence depths between the two stations was mainly caused by the total water input and soil texture. Secondly, the Hydrus-1D model was calibrated and validated using the long-term soil water content, soil matric potential, and recharge flux to groundwater. Calibration results showed that the model simulated the aforementioned processes well. Daily recharge amounts in 1992–2016 were simulated using the well-calibrated and validated Hydrus-1D model. Sub-recharge classification criteria with physical significance based on the daily recharge amounts were established, that is, the criterion for two precipitation events was no precipitation for seven consecutive days. Sub-precipitation events with irrigation inputs or no more than evapotranspiration were excluded. Sixty-seven and sixty-nine sub-precipitation events in 1992–2016 were separated at the LC and RZ stations, respectively. The average number of sub-precipitation events was no more than three per year. The sub-recharge amounts showed significant variability. In LC Station, the sub-recharge amounts were 0.1−421.7 mm, with an average of 22.2 mm and a standard deviation of 57.7 mm. In RZ Station, the sub-recharge amounts were 0.4−261.9 mm, with an average of 32.0 mm and a standard deviation of 65.1 mm. Finally, the relationship between the sub-recharge amount (sub-recharge coefficient) and the sub-precipitation amount (or total effective sub-water input, i.e., the sub-precipitation amount and water storage in the 0−200 cm soil profile) was quantitatively analyzed. There was a significant correlation (P<0.01) between the sub-recharge and sub-precipitation amounts (or total effective sub-water input), with a determination coefficient (R²) ranging from 0.801 to 0.962. However, there was no significant correlation between the sub-recharge coefficient and the sub-precipitation amount (or total effective sub-water input). This is because the sub-recharge coefficient cannot reflect the impact of the previous water storage in the soil profile on the recharge amount. The results of this study are theoretically significant for the evaluation of regional groundwater resources.