“This review explores the heart rate variability estimation technology based on facial videos, highlighting its non-invasive and real-time monitoring advantages in health monitoring and disease diagnosis. Deep learning technology, due to its powerful pattern recognition ability, can effectively extract complex visual features and process nonlinear physiological signals in HRV estimation, demonstrating significant advantages in improving estimation accuracy. This review aims to provide a comprehensive perspective on HRV estimation technology based on facial videos, providing important references for technological innovation and application expansion in academia and industry.”