Drowsy driving is a major threat to road safety, leading to countless accidents globally. This study introduces an innovative method to detect the driver drowsiness using convolutional neural networks (CNNs). By training the CNN model on an extensive facial image dataset, it can effectively recognize patterns linked to drowsiness. To boost accuracy, techniques such as facial landmark extraction and data augmentation are integrated. The proposed system provides a proactive approach to reduce the risks of drowsy driving, potentially enhancing road safety, and can also reduce road accident by enhancing driver awareness and response times. Future improvement can also be embedded to make it more accurate that may involve tracking the other signs of drowsiness like yawning and head position.