As an important part of urban development, the safety and quality of construction projects are directly related to the safety of people’s lives and property and the sustainable development of the city. With the development of construction technology, the scale and complexity of construction projects continue to increase, and traditional manual inspection methods are no longer able to meet the needs for efficient and accurate inspection. Therefore, this article uses advanced deep learning technology to automatically detect and diagnose construction engineering defects. This article focuses on the description and classification standards of defects in construction projects, as well as the methods of collecting and preprocessing defect data, and also analyzes the application of deep learning models in defect detection, and briefly discusses the subsequent model training and optimization strategies. Finally, two sets of simulation experiments were conducted and the results are as follows: the optimized construction engineering detection accuracy increased by 6.2% on average, and the detection efficiency increased by 17.35%.