The widespread application of Unmanned Aerial Vehicles (UAVs) is often constrained by their dependency on GPS, particularly in environments lacking reliable signals due to obstructions, indoor operations, and susceptibility to interference. This limitation underscores the critical need for GPS-denied navigation solutions, essential not only for indoor settings but also for complex environments like dense forests, tunnels, warehouses, and factories where UAVs play vital roles in surveillance and search and rescue operations. Achieving accurate and robust state estimation and control is pivotal for enabling autonomous navigation in these challenging contexts. Vision-based techniques such as Simultaneous Localization and Mapping (SLAM) and Visual Odometry (VO) have proven indispensable for precisely estimating pose and stabilizing UAVs during flight. This paper introduces a robust stereo vision-based Visual Inertial Odometry (VIO) algorithm tailored for precise waypoint navigation of Micro Aerial Vehicles (MAVs) in complex GPS-denied indoor environments. By leveraging VIO, our approach accurately estimates the MAV's current state and trajectory along cyclic paths, ensuring stable flight and precise waypoint attainment. Experimental results in indoor environments show that our MAV achieves autonomous navigation with an accuracy of 10-20 cm over a total distance of 30 meters.