Abstract: In order to study the dynamic characteristics of the cage, parametric modeling and analysis of angular contact ball bearings were implemented by using the secondary development of PCL (Patran Control Language) on the platform of finite element software MSC Patran/Ls-dyna. This simplified the pre-process and improved the modeling efficiency when the finite element software was applied to analyze the dynamic characteristics of bearings. Firstly, experiments were carried out by the research group’s paired bearing differential cage dynamic characteristics test device to validate the simulation results. The cage motion with the operating condition of the rotating inner ring and fixed outer ring was captured by the high-speed camera and its centroid trajectory was obtained by image processing technology. And the simulation results were verified through comparing with the literature results. There was good consistency in the shape of cage centroid trajectory and the trajectory radius between current simulation result and experimental or literature results, verifying the correctness of the current finite element model and the reliability of the simulation results. From this, the angular contact ball bearing with 7003C was selected to investigate the influence of different installation directions (horizontal and vertical installation direction) on the dynamic characteristics of the cage. This study was carried out from the following three aspects: the centroid trajectory of the cage in the radial plane, the displacement of the cage in the axial direction, and the maximum stress distribution of the cage. The results showed that due to gravity, the cage was easier to form a circular centroid trajectory for angular contact ball bearings with horizontal installation than vertical installation. When the bearing was installed vertically, the mass center of the cage oscillated irregularly and deviated from the center of the bearing due to the effect of gravity at a lower speed. However, at higher speeds, the cage was likely to form an approximate circular centroid trajectory. When the bearing was installed horizontally, the direction of gravity was consistent with the direction of the axial component of the force between the ball and the cage. Due to centrifugal force, the centroid trajectory at different speeds was approximately circular. The bearing with horizontal installation had a greater fluctuation range of the axial displacement than that with the vertical installation under different working conditions. When the bearing was installed horizontally, the gravity direction of the cage was along the axial direction, and the collision between the ball and the cage increased the resultant force of the cage in the axial direction, therefore, the axial vibration also increased. As the rotation speed increased, the fluctuation amplitude of the axial displacement in the two installation directions increased, while the fluctuation difference between the two installation methods decreased, suggesting that the installation direction had a greater impact on the movement of the cage at low speeds. Finally, the influence of different installation directions on the stress distribution was studied. Under low speed condition, the maximum stress of cage occurred at the lintel of cage for the vertical installation while at the side beam of cage for the horizontal installation. With high speed condition, the maximum stress of cage occurred at the lintel for both horizontal and vertical installation. What’s more, the vertical installation had a larger value of the maximum stress on the cage than horizontal installation at different speed condition.