This study employs Multi-Agent System technology to simulate knowledge services that support subject based self-directed learning in university library Future Learning Centers. By analyzing learner needs and designing collaborative agent pathways, a multi-agent service framework is constructed with functions including learning path design, resource recommendation, intelligent Q&A, and feedback evaluation. A simulation focusing on the "Economic History" discipline demonstrates that the system effectively integrates library resources, improves service automation, and enhances personalized support. The approach offers a practical and replicable solution for libraries with limited resources, providing insights for the development of Future Learning Centers.