Urban-rural integration in peri-urban areas of metropolitan cities faces many challenges, involving resource allocation, coordination of social structure, and improvement of infrastructure in the process of urbanization. Traditional urban-rural planning models can no longer cope with the rapidly changing needs, and new ideas and methods are urgently needed. The introduction of AI models provides data-driven support for urban-rural integration, which enables more accurate planning guidance and path realization. In this study, an intelligent planning framework is proposed to comprehensively assess the key elements of urban-rural integration planning by analyzing the regional characteristics and development trends of metropolitan peri-urban areas and combining the predictive capability of AI models. First, a multidimensional evaluation system for urban-rural integration is established through data collection and analysis, involving various aspects such as economic, spatial, and social integration dimensions. Second, the path realization model of urban-rural integration is constructed by combining hybrid machine learning algorithms, and the planning gets the implementation strategies in different development stages. The results show that the planning scheme based on the AI model can improve the efficiency of resource allocation, optimize the spatial layout of urban and rural areas, and effectively promote the sharing and flow of social resources. Finally, through simulation and field verification, it is concluded that intelligent planning not only improves the science of decision-making, but also provides a new path for the sustainable development of metropolitan peri-urban areas.