Abstract: Development of the plug-in hybrid electric vehicle (PHEV) with low carbon and environment friendly characteristics is an inevitable trend to achieve sustainable development in future. However, random charging behavior of large-scale PHEV may have serious impacts on the security and operation of distribution network. Firstly, a novel time-space dimension combined optimization scheme (TSCOS) is proposed in this paper, which investigates PHEV charging issue from time and geography viewpoints. With the connection between time dimension and space dimension, this approach manages to improve the overall system performance and minimize constrains on the PHEV owners’ charging behavior. Secondly, interval algorithm is used in the charging strategy to accommodate the random charging behavior and its uncertainty to distribution system analysis. Finally, case studies are presented to show the effectiveness of TSCOS with the IEEE 33 bus test system.