Abstract: How to accurately forecast the landslide has become one of the challenges of mining process. The paper describes briefly neural network and BP network model, and analyzes BP network's structural characteristics, parameter selection, data collection and processing, and construction of network model. A forecast model is established with the slope stability state in CNNC Jin'an Uranium Mine as the learning and training sample and the forecast sample. The stability analysis method based on BP neural network technology and its effectiveness are discussed. The real case computation shows that the forecast of slope stability by neural network method has achieved satisfactory results, providing a reference to the assessment of the similar slopes and the feasibility of applying the neural network in mine slope analysis assessment.

Key words: Uranium mine slope, Neural network, Forecast network model, Slope safety