In order to provide realistic ventilation flow rates, it is of great importance to know indoor pollutant concentration field quickly and efficiently resulting from any type of indoor-pollutant-source distribution, further facilitating the design and control of indoor ventilation systems for practical application. This work introduces the development of reduced-order ventilation models for transient pollutant dispersion. In particular, we focus on transients resulting from a step change in pollutant source distributions. We further focus on the decay problem. A reduced-order ventilation model is the solution for this decay problem, which is derived from a large coupled system of Ordinary Differential Equations (ODEs) for concentration that can be cast in terms of a matrix exponential, that is accurately represented with only a few dominant eigenmodes. Using a 2D ventilation case, dominant eigenmodes with their physical relevance and pollutant concentration results are presented. We find that the first 4 eigenmodes are sufficient to predict the pollutant concentration decay for the current test case. We also find that the complex eigenmodes play an important role in the indoor recirculation processes.