The use of prefabrication for buildings has been found to lead to benefits including improved efficiency in the construction processes, and reduced waste and environmental effects over a building's life cycle. Studies of buildings' life cycle carbon emissions (LCCa) have been conducted, but the unavailability of reliable data restricts the estimation of accurate LCCa of prefabricated buildings, making it difficult to compare the results using databases with different backgrounds. This paper aims to examine the differences among the existing databases for estimating prefabricated buildings' LCCa. In total, 17 databases focused on construction life cycle assessment (LCA) were analyzed and three databases were compared on prefabricated building component level. Facades, kitchens and staircases of a high-rise prefabricated building in Hong Kong were selected for detailed comparison. The results show that the carbon emissions calculated using "ecoinvent" and the "ETH-ESU 96" databases are similar while the results calculated using "IDEMAT 2001" are different. The differences among the three databases are mainly caused by three reasons: the lack of enough data on construction materials, the differences in the geographical representativeness of the material extraction and electricity generation, and the inconsistent LCA system boundaries used in these databases. Based on these findings, actions are suggested to improve the comparability of different databases, namely, developing a uniform communication system for LCA databases, enriching the open LCA databases, and achieving the consistency of LCA system boundaries. This paper provides a starting point to select a suitable database for prefabricated buildings' LCCa. It also presents an effective method to overcome the barriers to achieving comparability among different databases and benchmark the carbon emissions calculated using different databases.