Waste is an environmental issue that continues to escalate in line with population growth and human activities. One effective solution to mitigate the impact of waste is to
implement early sorting processes based on material types. This research designs and develops an automated waste sorting system utilizing Artificial Intelligence (AI) with the YOLOv11 algorithm to detect and classify waste objects into three main categories: organic, inorganic, and metal. The system is integrated with an Arduino microcontroller
to control servo actuators that push waste into the appropriate sorting bins. The testing results indicate that the YOLOv11L object detection model achieves a high level of
accuracy, with an average mAP50 of 95% and precise classification performance on the test dataset. Out of 32 trials, the system successfully sorted 28 objects into their correct bins, resulting in a sorting efficiency rate of 87.5%. However, the system encountered challenges when processing closely spaced or overlapping objects, leading to
misclassification due to unsynchronized actuator activations. The primary limitation lies in the absence of an object queue management mechanism, which hinders the system’s ability to prioritize sorting actions when two different types of objects enter the detection area simultaneously. For future development, it is recommended to implement a mechanical object flow control mechanism and integrate object tracking algorithms to enhance the system's precision and stability under real-world operational conditions.