This paper derives the geometric meaning of derivatives and gradients using normal vectors.This approach is not only intuitive but can also be directly extended to high-dimensional cases,offering a geometric interpretation for methods like gradient descent and other applications.