The co-occurrence network was established to explore the differences between the gut microbiome of patients with major depressive disorder (MDD) and healthy controls (HCs). Methods: Thirty-five MDD patients and 35 HCs were recruited, and their stool samples were collected for sequencing. Correlations between microbial species were calculated to construct the co-occurrence network of gut microbiome. Differences in network topology attributes and key nodes between groups were analyzed. Results: The analysis of gut microbiome co-occurrence networks showed that the microecology of the gut community was changed in MDD group. The decreased average degree and connectance of the network indicated the decreased interspecific interaction. Key members of the microecology were also found changed. The module hub for HC was Ruminococcus flavefaciens, and the connectors were Clostridium CAG 389 and Lachnospiraceae bacterium A2. The uncultured Blautia sp, Bacteroides mediterranensis and Parabolides distasonis were identified as the module hubs for MDD. Besides, the abundance of Ruminococcus flavefaciens and Clostridium CAG 389 both showed a significant correlation between clinical depression and anxiety scale scores. Conclusion: The co-occurrence network of gut microbiome in patients with MDD is different from healthy controls. To obtain more findings beyond pure abundance, co-occurrence network analysis needs to be further promoted and developed in researches related to gut microbiome.