In transition to a new era of space exploration, the question is what the next space logistics paradigm should be. The past studies on space logistics have been mainly focused on a “vehicle” perspective such as propulsive feasibility, cargo capacity constraints, manifesting strategies, and crew and vehicle demand, all assuming a predefined logistics network. Against this background, a graph-theoretic modeling approach to space-based resourceeconomy from a “network” perspective has been proposed. Built on this proposal, this paper continuously develops a graph-theoretic modeling framework and presents a multicommodity network flow formulation using flow transformation matrix. The proposed modeling method is demonstrated with a case study of cislunar architecture for human exploration of Mars. Mars capture orbit (MCO) is selected as a problem boundary node with a demand based on Mars Design Reference Architecture 5.0 and the problem is formulated as a linear programming (LP). The optimized network flow suggests that the resource depot should be located in Earth-Moon Lagrange point 2 (EML2), the orbital transfer vehicle should travel between geostationary transfer orbit (GTO) and EML2 like a pickup bus, and the Mars Transit Vehicle (MTV) should be injected into trans-Mars trajectory from EML2. It is found that utilization of lunar resources could reduce the total resource required to satisfy the same demand at MCO by up to 35.8%. Nomenclature