Abstract: A pressure forecasting method of submarine-borne missile cover system based on neural network evolved by genetic algorithm is proposed.Genetic algorithm is made completely self-adaptive depending on individual fitness.In addition,a new stochastic gene offset operation is used that improves the initial evolving velocity of the genetic algorithm.Training results showed that dummy function formed from optimum neural network has a good adaptability and can reflect the internal relation of samples.Using this method,pressure of submarine-borne missile cover system has been forecasted accurately.So the method can be used to forecast the pressure characteristics of submarine-borne missile cover system fastly and accurately.