Abstract: The logic regression model combining deterministic coefficient to solve the probability of landslide, the analysis of the slope state values are different, solving the probability of landslide is obtained, draw a state of absolute values of the slope increases, the probability of landslide is obtained with the sample input values is consistent, but as a forecast is, the greater the chance of error, in the process of practical application, considering the actual need to optimize the slope state values. Using logistic regression models and deterministic coefficient, with 30 samples slope as an example, the application of SPSS software for regression analysis, to solve the state values of different slope landslide probability. Will solve state values of different slope landslide probability results is analyzed by combining with safety factor, it is concluded that the state value of absolute value, the greater the slope, the probability of landslide is 0.00% and 100.00% to the landslide probability near the poles, and the probability of landslide is the smaller agv, shows that the landslide probability and sample input values is consistent, but used to predict the target of the slope landslide probability, there is the possibility of error, the greater the slope with samples, select the optimal value of absolute value of slope state of 4.