In this paper the suitability of Euclidean distance, similar disparity, similarity coefficient, correlation coefficient and best similarity coefficient used as a similarity measure in meteorological grid dataset is discussed. The results show that some special systems’ position and intensity cannot be expressed correctly by distance algorithm. But using similar disparity, similarity coefficient, correlation coefficient and best similarity coefficient, two meteorological grid datasets’ spatial distribution and the change trend are considered separately, so the similar ambient field can be found in the long history dataset, which can be applied in operation. In order to evaluate the similarity measures’ operational application, the similar ambient field corresponding to tropical cyclone (TC) track’s motion quantity is referenced in current source TC track forecast. The results show that 24 h moving direction forecast accuracy is better than that of 24 h track forecast and the similar disparity is better than other algorithm in track forecast, but the best similarity coefficient is better than other algorithm in moving direction forecast. By the query of similar ambient field, a similar TC ensemble corresponding to similar ambient field is found, which can be referenced in TC operational, forecast or research.