With the acceleration of urbanization, the complex interactive relationship between population growth and social network information dissemination has a profound impact on urban governance. However, existing studies lack systematic modeling of the dynamic coupling mechanism between the two, which makes it difficult for policymakers to quantitatively evaluate the feedback effect of information dissemination on population migration. To this end, this paper proposes a coupling modeling framework that integrates system dynamics and multi-agent simulation. First, this paper constructs a subsystem dynamics model of population growth based on urban economic, resource and environmental parameters. Secondly, a multi-agent simulation model is designed by combining the topological structure of social networks with user behavior data. Finally, a time-varying weight algorithm is introduced to realize cross-scale collaborative simulation of the dual system. Experiments show that information dissemination in social networks has an average lag effect of 5 days on population migration behavior; the policy intervention effect improvement rate is as high as 137%; the model shows excellent accuracy in predicting population change trends, with a goodness of fit R2 of 0.992. The research results verify the nonlinear synergistic relationship between the population system and the information system. The proposed model provides a feasible path and quantitative support for population regulation and policy response optimization in the context of smart cities.