Abstract: Landslide displacement prediction is a critical component of disaster prevention and mitigation. Rainfall infiltration directly affects the distribution of pore water pressure within the landslide mass, and the dynamic variation of pore water pressure significantly influences the shear strength of slip zone soils, making it a key factor in landslide initiation. To address the limitations of traditional landslide displacement prediction models that inadequately consider the coupling effects of pore water pressure, this paper proposes a novel displacement prediction method that incorporates pore water pressure. A one-dimensional diffusion model is used to simulate the changes in pore water pressure. Pore water pressure is introduced as a new influencing factor, and the optimal set of the influencing factors is determined using the maximum mutual information coefficient method. The selected influencing factors served as inputs to the displacement prediction model, while cumulative landslide displacement is used as the output. The method is validated using GNSS monitoring data from the Dangchuan landslide in Heifangtai, Yanguoxia Town, Yongjing County, Gansu Province, China. Results show a strong correlation between pore water pressure and landslide displacement. Incorporating normalized pore water pressure into the model significantly improves prediction accuracy, especially during periods of intense rainfall. The model achieves an RMSE of 2.9 mm, an MAE of 2.5 mm, and a coefficient of determination (R²) of 0.995, providing a promising approach for the accurate early warning of rainfall-induced landslides.