In view of the fact that the traditional model such as statistical model,deterministic model and mixed model in the dam safety forecast is insufficient,the paper combined the genetic algorithm(GA)with BP algorithm which based on error back propagation,the dam slope deformation prediction model of genetic optimization of neural networks(GA-BP model) is established.GA-BP model uses neural network misalignment mapping ability,network inference and forecast function,and the genetic algorithm global optimization characteristic to overcome BP algorithm is prone to local minimum.Using the model to forecast some dam's actual observation data,the forecasting result indicated that the GA-BP forecast model has the high precision,the quick convergence rate merit,and it is valuable to dam forecast aspect.

The multi-stage fuzzy sets theory is adopted to make a judging model on the damaged degree of environment because of coal mining. The model included five major reflecting factors. Meanwhile factor system and three stages judgment models are set up. Combined with the practice of Fuxin mining area, five surface subsidence areas are judged. The evaluated result is in according with the mine practice. So it provides a new practical method for evaluating the condition of geological hazard in mining subsidence area.