Accurate estrus detection and timely insemination are crucial for increasing productivity and economic outcomes in modern large-scale pig farming. Traditional methods of estrus identification, such as the back-pressure test, boar exposure, and physical inspection, are limited by high labor and time intensity, as well as human error. To precisely identify sow estrus. This study presents a deep learning model called YOLOv8n. The model was trained and validated using a dataset gotten from local sources and compared to Random Forest. Experimental data suggest that YOLOv8n obtained an accuracy of 97%, precision of 100%, recall of 95%, f1-score of 97%, and specificity of 100%. This enhanced model consistently distinguishes between estrus and non-estrus, beating the previous model. Validation confirms the model's robust performance in complicated contexts, which approaches expert-level accuracy in estrus identification. YOLOv8 enables consistent, continuous monitoring of estrus status in difficult situations and offers a fresh scientific approach to estrus identification for intensive pig farming.