Modern gripper systems of mechatronic modules still face significant challenges in grasping delicate, irregularly shaped, and deformable objects within uncertain working environments. Traditional rigid-structure grippers generally operate based on a simple open-close mechanism and lack adaptive tactile feedback systems. As a result, object slippage, excessive gripping force, and instability during manipulation processes often occur. Therefore, the demand for intelligent, adaptive, and economically efficient gripping elements is steadily increasing. This article discusses the improvement of a relatively low-cost adaptive robotic gripper system capable of detecting slippage in real time and intelligently adjusting gripping force using sensor integration and machine learning-based control algorithms. The proposed system includes the analysis of the operating principles of force-sensitive resistor (FSR) sensors and flex sensors to obtain real-time tactile and motion-related data during the gripping process. In the process of improving the gripping element of the mechatronic module, the signals obtained from these sensors are filtered and processed through feature extraction stages, after which machine learning classifiers are used to analyze slipping and non-slipping conditions. Subsequently, the gripping force is automatically adjusted through a PID-based adaptive control strategy.