探讨了采用遗传算法 (GA)进行旋转机组工作状态趋势预测的新途径 ,指出了适合度的选择是实现 GA算法趋势预测的关键。提出了 GA预测的方案 ,探讨了神经网络 GA预测、时间序列模型 GA预测 ,以及灰色模型 GA预测的适合度选择 ,提供了进行遗传算法预测的操作步骤。 This paper discusses a new way of forecasting trend of working status of rotating machines by genetic algorithm (GA), and points out that the choice of fitness is the key to forecasting trend of GA algorithm. The scheme of GA prediction is put forward. The selection of fitness for GA prediction of neural network, GA prediction of time series model and GA prediction of gray model are discussed, and the operation steps of genetic algorithm prediction are provided.