This study explores the application of advanced machine learning models in enhancing real-time object detection within daycare centers, focusing on ensuring the safety and security of children and staff. A comparative analysis was conducted on five prominent deep learning models: YOLOv8, Faster R-CNN, SSD, EfficientDet, and RetinaNet. These models were evaluated based on their accuracy, speed, and efficiency in detecting and tracking potential hazards and unauthorized individuals in daycare environments. A custom dataset consisting of over 20,000 annotated images and video frames was created, and each model was trained using transfer learning techniques. The study revealed that YOLOv8 outperformed the others in terms of speed and overall accuracy, making it ideal for real-time applications. However, models like Faster R-CNN and RetinaNet, despite their slower processing times, provided better accuracy in detecting rare events. The findings underscore the importance of selecting the appropriate model based on specific deployment needs, balancing between speed and accuracy.