ABSTRACT:   Open Source Software (OSS) is the software which grants access to its source code. A better understanding of how OSS functions may help developers take a more effective means for the system development. Studying software systems by the complex network theory can potentially provide useful insights into the diversity and success of OSS. However, research efforts on OSS concentrates on relationships among software systems. Little attention has been paid to operating systems themselves. In this paper, a data collection framework is proposed, and the distribution network of Linux operating systems is constructed. The network models distributions as nodes and dependencies among them as edges according to data crawled from DistroWatch. It is found that the distribution network is a scale-free and small world network similar to those identified in other fields. The inconsistency in share of distributions and targeted users is revealed and discussed. It is expected the constructed model can be a guide for distribution development in the future.