"数值微分法" can be translated as "numerical differentiation method" in English.
This term refers to a set of techniques used to approximate the derivative of a function numerically, especially when an analytical expression for the derivative is difficult or impossible to obtain.