It is well known that it often leads to large error to calculate the first order partial derivatives of physical quantities by the difference method from observation data with error. Especially in meteorology, observational wind has large error, and calculating vorticity from observational wind is very inaccurate. Therefore, one-dimensional numerical differentiation algorithm is applied to calculate vorticity from observational wind, and the result is compared with that of the central difference method. It turns out that the one-dimensional numerical differentiation algorithm is stable and feasible. Its accuracy is superior to the central difference method, and it has stronger recognizing ability for the smaller-scale weather systems.