The recent degradation and loss of mangroves jeopardize the health of coastal regions, aquatic biodiversity, people's livelihoods, and other ecosystem goods and services. While planting Rhizophora spp. has become a trend, the focus of mangrove conservation has now shifted to promoting biodiversity restoration, ecosystem functioning and services. Forecasting how the latter will be affected by conservation interventions and novel scenarios related to environmental change has thus been becoming increasingly important for effective mangrove restoration planning. Agent-based models (ABMs) describing mangrove regeneration and forest dynamics are well suited to this task because they simulate tree growth and physiological processes at the individual tree level while considering dispersal, competition and establishment that shape mangrove forest structure and diversity. We started this study by empirically linking natural regeneration, composition and functionality (NRCF) with different life cycle stages of mangrove trees. We reviewed the literature on existing mangrove forest ABMs to get an overview of how they address different aspects of NRCF. The results revealed that most models are so-called ingrowth models which focus on the later stages of the life cycle (saplings and trees). A few mechanistic models were found which describe explicitly natural regeneration (NR). They included key processes like seed production and propagule dispersal, but flowering and fruiting phases have been largely ignored. The existing mangrove forest ABMs implicitly address some ecosystem services such as carbon sequestration, while other services (e.g. coastal protection) have rarely been addressed by existing mangrove ABMs. We thus conceptualized these services and explained why an NR module can be a suitable base to describe and forecast these services. We then discussed the challenges of upgrading ABMs with such a module. Finally, we provided an outlook on how to address these challenges through incorporating ecological patterns into the design of the NR module.