In this paper, we take a mobile humanoid robot as the research object, and study the coordinated waist-neck motion strategy that can simulate human motion characteristics, so that the robot can accurately adjust the target to the center of its head camera’s field of view with the help of the coordinated waist-neck motion when it performs the framing task. The framing task is divided into two stages: target search and visual focus adjustment. For target search, a two-stage search strategy combining large-scale vision transfer and small-scale target search is proposed. For visual focus adjustment, joint comfort is taken as the goal and camera imaging is taken as the constraint to construct a waist-neck coordinated motion trajectory optimization function to achieve the purpose of adjusting the search target to the center of the field of view and avoiding the bad posture of the robot. Finally, using Coppeliasim as the simulation platform, the kitchen experimental scene is constructed, and through the complete framing experiments, the coherent and natural robot motion and accurate framing task can be realized, which verifies the effectiveness of the algorithm.