采用BP神经网络模型 ,以西北内陆河黑河流域莺落峡年平均出山地表流量为研究对象 ,对人工神经网络研究方法在干旱区环流径流预报中的应用进行了初步尝试 .结果表明该方法预报成功率较高 ,证实了人工神经网络方法应用于流量预报领域的可行性 ,并分析了该方法在预报过程中的优缺性 BP neural network model was used to study the mean annual surface flow of outflow from Yalekasia in the Heihe River basin of the inland river in Northwest China and the application of the artificial neural network method in the runoff and runoff forecasting in the arid area was studied. The success rate is high, which confirms the feasibility of artificial neural network method applied in the field of flow forecasting, and analyzes the advantages and disadvantages of the method in the forecasting process