Abstract: Rainfall is the main factor inducing slope deformation and instability, and the early warning and prediction of rainfall-triggered landslide has always been the core issue in the field of engineering. In this paper, the Monte Carlo method is introduced into the early warning and prediction of rainfall-triggered landslide. Firstly, based on the physical and mechanical parameters of rock and soil mass with normal distribution, the finite element numerical calculation model of slope is established, and the variation of slope stability coefficient under 9 different types of rainfall is analyzed. The results show that the incremental rainfall is particularly unfavorable to the stability of the slope, the uniform rainfall is the second, and the decreasing rainfallis the least. Secondly, the rainfall process is divided into antecedent and current rainfall, and the effective time of antecedent rainfall for current rainfall is 6 days. Finally, combined with the reliability theory, the failure probability P_f=10% is selected as the early warning index. By introducing the previous rainfall into the rainfall intensity-rainfall duration relation curve and taking it as the third coordinate axis, the curve is finally extended to the antecedent rainfall (A) - current rainfall (I) - rainfall duration (D) surface (A-I-D threshold surface). The research results have acertain guiding significance for the early warning and prediction of rainfall type slope.