Rainfall infiltration is a critical factor inducing slope instability, and simulating the infiltration process triggered by rainfall is essential for understanding slope failure and related geological hazards such as landslides. Based on indoor infiltration experiments conducted on loess soil columns, this study employs HYDRUS-1D to simulate the soil moisture infiltration process and FLAC3D to analyze the impact of rainfall infiltration on slope stability. The results indicate that the simulated soil moisture content values at various depths by the HYDRUS-1D model align well with the measured values over time, with a root mean square error (RMSE) ranging from 0.023 to 0.039 cm3/cm3 and a Nash-Sutcliffe efficiency (NSE) exceeding 0.7, demonstrating an excellent simulation performance. Additionally, the soil moisture content exhibits characteristics of prolonged response time and reduced amplitude of variation with increasing soil depth. Under non-ponding conditions, the measured values of soil moisture infiltration are highly consistent with the simulated ones of the wetting front movement, with a correlation coefficient of 0.997. The model’s accuracy in simulating the timing and velocity of wetting front movement improves progressively with soil depth, resulting in a significant water flow lag effect. Analysis using the FLAC3D model reveals that increased rainfall infiltration depth causes the shear zone within the slope to expand from the toe to the crest, leading to gradual slope instability. The safety factor decreases from 2.46 under natural conditions to 0.72 when the infiltration depth reaches 200 cm, indicating a significant impact of rainfall infiltration depth on slope stability. These findings provide important insights into the changes in moisture fields in loess slopes and the prediction of landslides, contributing to a better understanding and prevention of geological hazards.