In the near future, the waste management sector is expected to reduce substantially the adverse effects of garbage on the environment. However, the increasing complexity of the current waste management systems makes the optimization of the waste management strategies and policies challenging. For this reason, waste prevention is the most desirable goal to achieve. Despite this, low levels of household recycling represent the key factor that complicates the current scenario. Keeping this in mind, the present work investigates the determinants of recycling behavior through the development of an agent-based model. Particularly, we examined what would induce households to increase the probability to engage in recycling behaviors on the base of the individual attitude and sensitivity to social norms. The Theory of Planned Behavior (TPB) has been implemented as agents' cognitive model in environmental studies with the aim to predict recycling outcomes. Furthermore, in order to increase the realism of the simulation and the adherence of the model with the theory, we followed two strategies: firstly, we used real data to model a city district (Diong, Internship Report: Integrated Waste Management in Kaohsiung City, 2012). Secondly, we made use of the coefficients of the structural equation model presented in the work by Chu and Chiu (J Appl Soc Psychol 33(3):604-626, 2003) to build the agents' cognitive model. As a whole, the results are in line with literature on descriptive social norms. Furthermore, the results indicate that the introduction of descriptive social norms represents a valuable strategy for public policies to improve household recycling: however, injunctive social norms are needed first.