Abstract: The research on long-term variability of ocean waves is crucial for human life and production. However, reanalysis data often underestimate wave heights in offshore region and high sea conditions, and may have potential negative impacts on subsequent applications. To address this issue, this paper integrates the swell height data in the past 40 yeras from the International Comprehensive Ocean-Atmosphere Data Set with those from ERA5 reanalysis using the Data-Interpolating Variational Analysis in n dimensions algorithm. The study compares the original reanalysis wave heights with the fused data with the methods such as K-fold cross-validation, satellite data validation and spatial-temporal distribution analysis. The results indicate that the fusion method effectively enhances the swell wave heights in long-term data, making the fused swell wave heights closer to actual observations in high sea conditions. Furthermore, while preserving the major spatiotemporal variation characteristics of the original reanalysis data, the fusion method reveals additional subtle variation modes, which facilitates subsequent work in marine disastermitigation, offshore engineering, and other related fields.