Some methods of fuzzy sels are applied to the analysis of water masses in the north area of the South China Sea. In distinguishing water masses with the fuzzy clustering method, the authors propose a fuzzy similar relation based on Euclidean distance, and suggest a method for determining the number of water masses by the F-test, and further adjust rationally the boundaries of water masses and the membership relation of samples by the elastic classification method. Eight water masses are distinguished in the area, i.e., Nearshore Diluted Water F, Nearshore Mixed Water M, Warm Surface Water WS, Surface Water S, Surface-Subsurface Mixed Water SU, Subsurface Water U, Subsurface-Intermediate Mixed Water UI and Intermediate Water I. These point-sets of various water masseson the (?)-S diagram may be summed up in three types, that is, the deformed elliptical convexpoint-set, the half-boundless convex point-set and the point-set with single-side concave. For them, we have proposed methods of making their member Some methods of fuzzy sels are applied to the analysis of water masses in the north area of ​​the South China Sea. In distinguishing water masses with the fuzzy clustering method, the authors propose a fuzzy similar relation based on Euclidean distance, and suggest a method for determining the number of water masses by the F-test, and further adjust rationally the boundaries of water masses and the membership relation of samples by the elastic classification method. Eight water masses are distinguished in the area, ie, Nearshore Diluted Water F, Nearshore Mixed Water M, Warm Surface Water WS, Surface Water S, Surface-Subsurface Mixed Water SU, Subsurface Water U, Subsurface-Intermediate Mixed Water UI and Intermediate Water I. These point-sets of various water masseson the (?) - S diagram we have summed up in three types, that is, the deformed elliptical convexpoint-set, the half-boundless convex point-set and the point-set with single-side concave. For them, we have proposed methods of m aking their member