To clarify the theoretical foundation of robotic vision guidance technology, we first introduce the basic architecture of the robotic vision guidance system from a general perspective. By establishing the mathematical model of robotic vision imaging and spatial target measurement, we derive the three-dimensional measurement theory of different vision measurement systems, such as monocular, binocular, multi-ocular, and structured-light systems. The target pose estimation theory based on feature points and point cloud registration is introduced, followed by the forward and inverse kinematics of robots. The joint robots are used as an example to derive the forward and inverse kinematics analysis method, which is verified through Simulink and ADAMS joint simulation.