© 2021 Elsevier LtdThis study aims to determine long-term spatial variability of wind and wave characteristics over the Sea of Marmara. For this purpose, we produced four different wind and wave datasets for a period of 40 years (1979–2018) using the calibrated and default-setting SWAN models forced with both CFSR and ERA-Interim winds and inter-compared. The performance of these datasets is examined against the wave height measurements of 2015 at the Silivri buoy station and 27-year (1991–2018) satellites data (ENVISAT, ERS-1, ERS-2, JASON-1, JASON-2, JASON-3, SENTINEL-3A, SENTINEL-3B, TOPEX). Then, for wind and wave climate analysis, the spatial distributions of the wind and wave characteristics over the study area are determined in terms of seasonal and annual averages. 40-year averages of 95th and 99th percentile significant wave heights, probabilities of being larger than significant wave heights exceeding a critical value, and 40-year maximum significant wave heights are spatially analyzed. Finally, the calibrated SWAN model using CFSR winds presents the most accurate results. Therefore, the effects of different model inputs and settings on the model results are examined by creating spatial distributions of the biases of the other three model results and comparing them to those of the calibrated model results using CFSR winds. The CFSR forced and calibrated model predicts the largest average wind speed and average wave height in the central regions of the Sea of Marmara in comparison with the other three models. The shores of Kapidag Peninsula and Marmara Island are exposed to be with the highest wave conditions.