This study constructed datasets of Significant Wave Height (SWH) from multi-source satellite altimeters using the linear revision method and the probability density function method(PDF) in the China offshore waters, compared the results from these two methods, tested the consistency of these fused climate datasets, and analyzed the inter-decadal trend of the fused SWH datasets. The results show that: (1) Cross-calibration results using both linear and PDF methods reflect consistent spatial distribution characteristics of SWH; (2) Relative to the reference data, in the ranges of extreme high or extreme low values of the SWH, the results from the PDF method show lesser bias than that from the linear revision method, and are closer to the reference data; (3) The linear trend from 1992 to 2024 indicates that compared with the linear method, the significance level of the results from the PDF method cross-validation is higher, with a negative interdecadal trend of -0.05 m/10 years in China offshore areas, which means the value is closer to that of the reference data. This study provides a new method for constructing of multi-source satellite remote sensing climate datasets.