Human agency is an essential determinant of the dynamics of agroecosystems. However, the manner in which agency is represented within different approaches to agroecosystem modeling is largely contingent on the scales of analysis and the conceptualization of the system of interest. While appropriate at times, narrow conceptualizations of agroecosystems can preclude consideration for how agency manifests at different scales, thereby marginalizing processes, feedbacks, and constraints that would otherwise affect model results. Modifications to the existing modeling toolkit may therefore enable more holistic representations of human agency. Model integration can assist with the development of multi-scale agroecosystem modeling frameworks that capture different aspects of agency. In addition, expanding the use of socioeconomic scenarios and stakeholder participation can assist in explicitly defining context-dependent elements of scale and agency. Finally, such approaches, however, should be accompanied by greater recognition of the meta agency of model users and the need for more critical evaluation of model selection and application.