Insufficient understanding of complex Arctic cloud properties introduced large errors in estimating radiant energy balance parameters at the regional and global scales. Comprehensive and reliable cloud information is necessary for improving the accuracy of flux inversion. This study evaluated daytime cloud fraction (CF) uncertainties from 16 available satellite products and estimated the spatio-temporal distributions of Arctic daytime CF during 2000–2019. Our results show that the differences among multiple products had significant temporal and spatial heterogeneities. Temporally, the maximum and minimum inter-product discrepancies occurred in April and the summer months, respectively. Spatially, the largest uncertainties were seen over Greenland. Substantial inconsistency also occurred on the Central and Pacific sides of the Arctic Ocean. The active satellite product tended to capture more clouds in these two regions. We found that the inconsistencies caused by different algorithms were smaller than those caused by the sensor itself, with ±2% and ±5% for MODIS-based datasets and ±6%, ±15% for AVHRR-based datasets. The annual average daytime CF in sunlit months was 70.9 ±2.93% and increased over the Arctic during study periods. These upward trends might cool the Arctic by approximately 0.05−0.5 W/m2/decade. In terms of the spatio-temporal distributions, the CF over the ocean is higher than that over the land, and the former increased significantly while the latter decreased; the CF trends of most products are positive in June and July but are opposite in other months. From this study, the findings based on multiple products would be more robust than that based on a single or few datasets.