简要介绍了两种智能化自适应模式识别方法预测工作面顶板来压的实施途径，即用灰色等维拓扑预测方法对来压作近、中期预报；采用人工神经网络方法实时预报顶板来压特征。针对煤矿特点首次提高出两种灰色等维建模途径及建立动态变化的模型群进行来压预测的技术思路。按这两种智能预测方法开发了微机软件，现场使用表明，预测预报准确度高、实用性强。 This paper briefly introduces two intelligent ways of adaptive pattern recognition to forecast the pressure of the roof, that is, using the gray isometric topological prediction method to forecast the near-to-mid-range pressure and the artificial neural network . For coal and mines for the first time to improve the characteristics of two gray-dimensional modeling approach and establish a dynamic model group to pressure prediction technology. According to these two kinds of intelligent prediction methods, the microcomputer software was developed. The field application shows that the prediction and forecasting accuracy is high and the practicality is strong.