Monitoring driver’s postures has extensive applications. The postural information could be used for the development of smart airbags, for detecting possible fatigue in long travel and for recognizing activities which may determine if the driver has enough time to take over the control in an intelligent vehicle when encountering hazardous situations. Microsoft Kinect is one of the best candidates for monitoring driver’s posture thanks to its innovative feature of real time motion capture without use of markers and its low cost. However, when body parts are partially occluded, the accuracy of Kinect data will drop markedly. Inspired by previous researches, the present work focused on testing a data driven approach for improving driver’s upper body movement reconstruction with a Kinect camera. Firstly, we organize a database of accurately captured driver poses from different motion clips with filtered structures, including a global motion graph and multiple local motion graphs. The reliability of individual skeletal joint from Kinect posture is objectively evaluated to work with weighted KNN to extract similar postures from the prior database. We then apply PCA to create an intermediary posture space where Simulated Annealing Algorithm is utilized to synthesize natural and kinematically valid postures. Kalman Filter is employed to help select the perspective posture candidates as well as to ensure temporal motion continuity across frames. The preliminary results show that the implemented method could remarkably improve the quality of reconstructed driver’s postures from Kinect and the framework has a great potential to be used to monitor the driver.