Autofocus technology plays an important role in the field of infrared thermal imager monitoring. At present, there exist some problems with infrared auto focusing technology, such as low success rate, complex architecture and low focusing speed. Therefore, this study proposes an auto focusing technology of infrared lens based on FPGA, which realizes the functions of infrared image processing, display, and auto focusing with a single FPGA. In view of the common characteristics of vertical stripe noise and random noise in infrared images, this study improves and optimizes the infrared definition evaluation algorithm and mountain climbing algorithm in the focusing process. The experimental results show that the algorithm and implementation method proposed in this study can help infocusing on the infrared lens remarkably. Meanwhile, the proposed method has characteristics such as high integration, fast focusing speed, and high success rate, and thus has wide application prospects.