Abstract: Slope engineering is usually designed under the shortage of test data. So the design method, how to integrate all aspects of the collection of information using the statistical theory to obtain engineering design parameters, is of great theoretical and practical significance. Based on the experimental data of two faults of a slope project, in accordance with the large number of measured data analysis results of similar projects, the rock classification method is used to determine the internal friction angle of the fault, and the back analysis method and the least-square fitting method with consideration of the residuals are used to determine the cohesion of fault; the anti-slide piles are the lateral force system, the correlation of the four failure modes of the anti-slide piles is considered, and the back analysis method is used to determine the statistical parameters of resistance of the anti-slide piles; and according to the (norm) recommended coefficient transmission method, the function of random variables of slope sliding is determined, and the discretization dimension reduction method is utilized to obtain the reliability index of slope stability. The results show that using the back analysis method to modify the parameter can decrease the uncertainty and improve the reliability of engineering design, thus providing a feasible design method using reliability in small sample data. The effect of anchor anti-slide pile reinforcement slope is remarkable.