Based on the quality data of a certain satellite control subsystem from 2009 to 2020， a technical risk level prediction model for the on-orbit satellite control subsystem is proposed based on BP neural network. The risk assessment system of the on-orbit satellite control subsystem is established by mining historical data. After eliminating the redundant information in the risk source， the important information is discretized and trained by the BP neural network. Finally， the risk level prediction model for the on-orbit satellite control subsystem is obtained. The experimental results show that the model can accurately and effectively predict risk levels for the on-orbit satellite control subsystem， and provide support for on-orbit satellite risk control.