Objective To effectively evaluate potential risks in intelligent manufacturing systems and improve the accuracy and efficiency of risk management， a risk factor evaluation model for intelligent manufacturing systems based on non-dominated sorting genetic algorithm II （NSGA-II） optimization is proposed.Method A risk factor evaluation model is established for intelligent manufacturing systems based on principles such as integrity， operability， and comparability. The initial evaluation index system is reduced to improve evaluation efficiency. Core indicator data is used as the input source for optimizing the neural network， which is deeply optimized by integrating the NSGA-II with elite strategies. The minimum mean square error of training data is set to enhance fitting ability， and the sum of squared weights is minimized to promote model generalization performance. NSGA-II optimized neural networks are utilized to achieve accurate and efficient evaluation of risk factors in intelligent manufacturing systems.Result The logarithmic loss of evaluation is less than 0.1 in the risk assessment of six intelligent manufacturing enterprises， demonstrating the effectiveness of this method.Conclusion This method can quickly respond to the risk changes of intelligent manufacturing enterprises and can provide timely and accurate risk assessment results for decision-makers. It not only improves the accuracy and efficiency of risk assessment， but also has certain practical value， which helps intelligent manufacturing enterprises to better deal with risks.