Chicken population counting on farms is still conducted manually by barn workers, which is time-consuming and prone to errors due to the chickens’ active movements and instances in which chickens obstruct each other’s visibility (occlusion). This study aims to design and develop a real-time chicken counting system based on computer vision using the YOLOv8-Nano algorithm on a Raspberry Pi 4. The system integrates a webcam as the image capture device, a Raspberry Pi 4 as the main processor, and a monitor to display detection results in the form of bounding boxes and the number of chickens in real time. Testing was conducted based on chicken density, lighting conditions, observation time, and chicken conditions. The test results showed that the system achieved 100% accuracy at low density, 95% at medium density, and 93.33% at high density. In the lighting tests, the system achieved 100% accuracy under bright conditions, 95% under moderate conditions, and 90% under dim conditions. The system also operated stably at various observation times and under conditions involving both live and dead chickens. The results indicate that the developed system can facilitate a faster, more efficient, and more accurate chicken counting process and serve as an alternative to manual counting methods.