Autonomous mobility and multi-agent systems have increased rapidly in business and personal use over the past few years such as mobile robots and autonomous vehicles. This study designs and develops a prototype of a multi-agent autonomous mobility control (MAMC) system that operates heterogeneous and autonomous agents with task assignments and mobility planning. The information on agent locations, collision avoidance, path planning, and action status can be synchronized to optimize integrated multi-agent mobility and operation. The system employs the Model–View–Controller (MVC) software design mechanism to facilitate backend information and data workflow. In the MVC design, the Model module represents the database of the mobility control system, the View module represents the user interfaces to visualize the map and locations of the autonomous agents, and the Controller module accepts input, processes the developed algorithms and converts it to commands for the Model or View module. Furthermore, a bidding-based task assignment algorithm is designed to assign tasks to multiple autonomous agents under the MAMC system, with value-based bidding considering the task costs including time and distance. Experimental testing is conducted in a Robot Operating System (ROS) simulation environment with a three-dimensional environment. Results indicate the capability and effectiveness of the designed MAMC for managing mobility controls, task assignments and path planning of autonomous agents.