Water transfer projects, while optimizing water resource allocation, have gradually developed into artificial ecosystems after long-term operation. The structures and assembly mechanisms of biological communities in such artificial ecosystems often exhibit characteristics that differ from those in natural river systems. This study focused on the East River Water Source Project, a strategic inter-basin water transfer project in Shenzhen, using bacteria as indicator species. Through bacterial sampling and high-throughput sequencing, the study investigated the variation patterns and assembly mechanisms of bacterial communities along the project. The findings are crucial not only for the water quality safety of the project's water supply but also for understanding its ecological implications. The study found that bacterial diversity initially decreased and then increased from the intake to the endpoint of the project. Significant shifts in diversity occurred at the pump stations, indicating that the changes in bacterial communities were related to the operation of the pump stations. Dividing the entire project into four sections based on the location of these pump stations, significant differences in bacterial community structures were observed among sections. The community differences were primarily dominated by conditionally rare taxa. Given the relatively stable internal environment of the project, the bacterial phenotypes across the sections were similar. Only aerobic, facultative anaerobic, and Gram-negative bacteria showed significant differences among the four sections, with aerobic bacteria being the dominant group throughout the project. Based on null model analysis of bacterial community assembly mechanisms, it was found that the bacterial community assembly was mainly governed by stochastic processes in the project. In stochastic processes, homogeneous dispersal was predominant, and dispersal limitation was observed only at the end of the project. Deterministic processes were only driven by heterogeneous selection, which gradually weakened along the project. This suggests that bacterial transport and dispersion caused by water transfer play a key role in the assembly of bacterial communities within the project. The random forest model analyzing the relationship between environmental factors and bacterial diversity also indicated that environmental factors had limited explanatory power in the model. Only golden mussel density and sulfate concentration have a significant impact on bacterial diversity. This suggested that the deterministic processes caused by environmental selection were weak. This study provides valuable insights into the evolution and adaptive mechanisms of artificial ecosystems in water transfer projects. It also offers a foundation for improving the management of water quality within such projects, highlighting the importance of considering microbial community dynamics in future ecological assessments.