Autonomous robotic assemble is the most promising approach to construct ultra-large space structures in the future. Ultra-large space structures usually is consisted of multiple modules, which requires the robot to perform repetitive tasks such as capturing, installing, and crawling on the flexible structure. In addition, the structure is affected by space environmental forces during the assembly process and experiences the growth in configuration and changes for parameters, which makes its dynamic behaviors very complicated. In order to study the assembly process of ultra-large multi-module structures in space, a simulation framework is proposed for the orbit-attitude-structure coupled dynamics, planning, and control. Firstly, natural coordinate formulation and absolute nodal coordinate formulation are used to establish an orbit-attitude-structure coupled model for the main structure, space robot, and assembly modules. The Kelvin-Voigt linear spring-damper model is employed to describe the contact and collision for the robot's gripper and the module's docking mechanism. Then, the motion planning, trajectory planning and joint trajectory tracking control of the robot are studied. Finally, dynamic simulations of the assembly process are carried out with different assembly attitude angles and attitude control schemes. Simulation results reveal that the main structure may experience significant orbital drift during assembly (depending on the assembly attitude) due to the change of the center of mass and the orbital difference between the main structure and assembly module. When the structural modules are assembled to the main structure in the radial direction, the semi-major axis and orbital eccentricity of the system will increase greatly. In contrast, the semi-major axis and orbital eccentricity remains unchanged when the structural modules are assembled in the tangential direction. In addition, even if the gravity-gradient-stability attitude is adopted, both attitude control and structural vibration control are necessary to reduce the vibration amplitude of the structures, minimize collision risk between the robot and structure, and improve assembly accuracy and assembly efficiency.