First-generation anti-tank guided missiles (ATGMs) mainly rely on the skills of operators to guide them in hitting targets, thus making them inaccurate. However, third-generation ATGMs are affected by electronic disruption and are extremely expensive. For example, the accurate range for anti-tank RPGs is 100 m, while the range of the ATGM Javelin is just 4 km (Lightweight CLU) and the cost is US$249,700 (Lightweight CLU only). However, the range of our system depends on the camera range used, which can exceed 50 km on flat ground, and the cost is a few thousand dollars. In this paper, an object detection model was used to recognize a targeted tank, and the fundamental matrix and triangulation approaches were utilized to determine the location of the target tank. Thus, the produced control circuit is smart, autonomous, accurate, with a wide range, and quite inexpensive. The system was installed on an anti-tank missile, tested indoors, and worked correctly. Using object detection models and computer vision technologies in the weapon industry led to the production of an intelligent weapon that is smart, has more range, and is autonomous.