With the rapid development of China’s economy, the field of construction engineering is also making continuous progress. Safety management has become more important in the construction industry, so improving the quality and level of construction has become a top priority. AI is the product of new technologies such as intelligence, information, automation and networking. This paper will start from the construction status of the project, combined with relevant theories to analyze the problems faced by the intelligent system and the solution measures. By using computer-aided means to prevent and deal with construction engineering accidents, and applying computer simulation technology to safety supervision work in practical operation, the construction quality and efficiency are improved. Then, this paper tests the performance of the safety supervision system, and the test results show that the alarm frequency data covers the number of data points at 100 time points, obtained by random generation, and the value range is between 50 and 100. These data points show a trend that reflects the alarm situation of the safety regulatory system. Before it is optimized by intelligent algorithm, the response time gradually decreases and shows a decreasing trend. However, the optimized response time starts at 0.8 and remains relatively low. Over time, the response time is further reduced. This shows that after optimization, the response time of the security alert system has been significantly improved, and it is able to respond more quickly.