Abstract: In this paper, by applying the conception of genetic algorithm into the parameters turning process of I/O schedule strategy, it takes strategy parameters adapting the storage system according to the periodical statistic of I/O operation, and got a group of strategy parameters, which have the best adaption in real-time storage environment, and makes the I/O schedule strategy have a better performance. In the new way of tuning strategy parameters, it makes the strategy parameters resuming its convergence in a short time. The experimental results show that, in the storage system which is running long term and have a heavy load, the efficiency of computer's I/O operation which applies this paper's strategy has an obvious performance promotion than applying others.