Most global road accidents are mainly caused by abnormal driving behaviors such as reckless driving, distraction, mobile phone usage, adjusting radios, or other vehicle systems. Driving while tired or drowsy can directly lead to accidents due to these abnormal driving behaviors. Additionally, this can result in property damage and pose risks to oneself and others, causing injuries or fatalities. Abnormal driving behaviors can be identified through facial expressions, yawning, blinking excessively, or looking elsewhere. Signs of fatigue include excessive blinking. To analyze factors contributing to road accidents, researchers have proposed the Unified Driver Attention and Fatigue Detection (UDAF) algorithm. This algorithm aims to enhance the efficiency of detecting abnormal driving behaviors by identifying gaze direction and blinking using Virtual Geometry Group (VGG16) and Long Short-Term Memory (LSTM). In this study, experiments were conducted using images and videos from front-facing cameras. The research achieved an average accuracy rate of 93.12% in detecting gaze direction and blinking through the UDAF algorithm.