Mobile robots have received much of attention in the last three decades due to their very high potential of applications such as smart logistics, exploration, and intelligent services. One of important functions of mobile robots is the navigation in which robot must know their location, the maps of environment and perform path planning with obstacle avoidance. In this work, we introduce an improvement of control algorithm for mobile robot using multi-layer sensor fusion toward the target of efficient obstacle avoidance. Based on our method, we used three layers of sensors arranging in three height-different planes of robot鈥檚 housing for sensor fusion. A control algorithm, which is extended from the so-called bubble rebound algorithm and uses signal from sensor system, was proposed. Experimental implementation on a mobile robot, named EAI, shows that our algorithm can control the robot to navigate and avoid obstacles much efficiently, in which obstacles in forms of different shapes and height can also be avoided. A high repeatability and stability of the algorithm is obtained.