Life-cycle assets management (LCAM) is promising for improving power system efficiency and reliabilities, lowering life-cycle costs of smart grid. This paper, after investigating and redesigning traditional assets management processes with integration-centered optimization, puts forward a life-cycle assets management system model with multi-agent style, in which life cycle analysis (LCA) can be executed with collected and retrieved data from heterogeneous agents in complex multi-level environment. Related critical issues are also discussed about system modeling and implementation. Finally, technical solutions are proposed about agent communication mechanisms and information security. Research results can be extended to enhance information capacities in related fields.