Abstract: With the development of new power systems, the penetration of renewable energy generation keeps rising, and distributed photovoltaics (PV) are connected to the grid on a large scale, triggering a series of problems such as voltage violations and transient voltage instability. Traditional local voltage control plays a crucial role in local regulation due to its rapid response capability, while centralized control demonstrates superior potential for coordinated regulation through global optimization. However, facing the drastically increased system complexity caused by high-penetration of renewable energy integration, local control fails to achieve global coordination, and centralized control suffers from excessive computational dimensionality and heavy communication burdens. By contrast, cluster-based partition voltage control effectively balances resource utilization efficiency and control feasibility. In view of this, this paper summarizes the existing partitioning index systems. From two dimensions including static physical connection and dynamic regulation sensitivity, it analyzes the adaptability differences of indices such as electrical distance and voltage sensitivity to the spatiotemporal fluctuation characteristics of sources and loads. Secondly, three mainstream algorithms, namely traditional clustering, community detection and intelligent optimization, are compared to reveal the evolution trend of partitioning methods from static topological rigid partitioning to dynamic sequential elastic aggregation. Furthermore, voltage control strategies across different time scales are reviewed from the perspectives of cluster autonomy and multi-cluster coordination. Finally, based on a systematic review of the existing technical frameworks, future research directions including data-driven control, heterogeneous resource aggregation and resilient defense are further discussed.