The application of UAV oblique photogrammetry technology and deep learning target recognition algorithm in the construction measurement of super-high building bridge pylons is studied. A rapid measurement technology for key points of bridge pylons based on deep learning is proposed. UAV oblique photography is used, and the three-dimensional model is rebuilt by Agisoft Metashape software. Combined with the YOLOv5 deep learning algorithm, the target points of pictures with geographic information are automatically recognized to automatically output the accurate geographic location information of key points of bridge pylons. Taking the piers in the construction process of Lanyuan Yellow River Bridge as an example, the feasibility and accuracy of the rapid measurement technology for key points are verified. The results show that this method can significantly improve the measurement speed and accuracy without affecting the construction quality. Through the combined use of UAV and deep learning algorithm, the high-precision rapid measurement of the complex structure of bridge pylons can be realized, and the geographic information coordinate values of the key points of the structure can be automatically output, which can meet the high-precision engineering requirements within an error of 1 cm.