This study investigates large-scale robotic systems depending on Multi-Target Path Planning (MTPP), where mobile agents should plan for their destination without collision and reach to the specified objective places. They restrict the range of real-world domains where they are applicable by using nonholonomic car-like agents instead of holonomic ones. This essay provides a multi-agent path finding mathematical formalization for the problem with known and unknown environments. Conflict-Based Search presents a unique hierarchical search-based solution. Omnidirectional multi-agent path planning has been proposed. We have modified our algorithm to be able to plan for an omnidirectional multi-agent mobile robot for multi-target in a warehouse environment by applying agents to reach their goals without conflict with the obstacles inside the environment. A comparative study has been made with related works, our work shows better performance in time, cost, and path length.