For many years, optimal global pathway design has been a fascinating field of study that keeps producing innovative methods. A robot may be trapped in local minima without an appropriate global path. This paper aims to navigate a custom robot in a Robot Operating System (ROS) environment using an improved A* algorithm. This paper proposes to solve several issues including smooth turns and local minima problems for a mobile robot in a complex scenario using an improved A* algorithm. To achieve results, the custom robot is designed, developing packages and creating launch files for the gazebo and rviz. Packages are added for Simultaneous Localization and Mapping (SLAM), Adaptive Monte Carlo Localization (AMCL), and A* Algorithm. Simulation is carried out using the Gazebo and Rviz simulators. The results of the simulation demonstrate that the improved A* algorithm can successfully maintain the intended path clear of obstacles, lower the robot's risk of collision, lower the frequency of the robot entering a local minimum, and raise the robot's stability.