In recent years, the aging of skilled workers and the shortage of young technicians have become serious problem in the manufacturing fields. Also, the international occupational health and safety standards are being developed, so it is becoming increasingly important to develop skills that take into account safety and physical load risks such as posture management and muscle load reduction during work in addition to work speed and accuracy. In this paper, we propose a pneumatic artificial muscle drive to support the optimization of work movement. We carry out a cross-modal movement safety analysis using image data of three-dimensional skeleton and documented work guideline data. The proposed system provides motion optimization support according to worker individual characteristics and has a mechanism that enables visualization and evaluation of the tendency to comply with rules during work. Experimental results show that the cross-modal analysis enables us to understand the trend of rule compliance and the pneumatic actuator supports motion stabilization adapted to specific rules. We plan to apply the proposed system in industrial and educational fields as a method to support skill retention and improve work safety.