The predominance and economic significance of small and medium-sized enterprises (SMEs) meanswidespread disruption to their operations can have severe financial consequences for a nation’s economy. For instance, the 2007 summer floods in the UK caused damage estimated at £2.3 billion and a significant proportion of this damage was sustained by SMEs, which, in 2018, represented 99.9% (5.7 million) of all UK businesses. Thus, due to the importance of this sizeof business to developed nations’ economies, this research seeks to develop an agent-based modelling and simulation (ABMS) framework to investigate manufacturing and retail SMEs’ immediate response to and short-term recovery from flooding. This ABMS framework consists of three main components: (i) flood-event simulation; (ii) a modelled geographical environment (MGE); and (iii) agent-based modelling and simulation. First, flood-event simulation represents the input (inundation) data that feeds into the MGE component, which provides a common platform for the flood-event simulation and the agent-based modelling and simulation. Specifically, the MGE combines Ordnance Survey (OS) MasterMap® data (i.e. Topography, Integrated Transport Network (ITN), and AddressBase Plus layers) with the input from the flood-event simulation to enable the identification of the flooded and non-flooded organisations. Consequently, a proportion of these identified organisations, and particularly SMEs from manufacturing and retail sectors, are modelled as autonomous agents and simulated over three stages, namely pre-flood, flood (no operation), and post-flood. In addition, SME-related organisations (e.g. suppliers, customers, electrical service providers, plumbing service providers, IT service providers, cleaning service providers, the Environment Agency (EA), mutual aid partners, refurbishment companies, and insurance companies) are also modelled as autonomous agents and simulated to support SMEs’ operation and/or recovery. In terms of SMEs, manufacturing and retail SMEs are modelled as autonomous agents with different sets of attributes (e.g. static and dynamic), behaviours (e.g. pre- and post-flood), and pre-defined precautionary measures (e.g. flood resistance, flood resilience, and businesscontinuity/risk management). Production level and service capability are computed at each simulation tick for every manufacturing and retail SME modelled, respectively, to determine their performance throughout the simulation. In terms of simulation, two real-life case studies of UK flood events in 2007 have been considered: (a) the Lower Don Valley of Sheffield, south Yorkshire, and (b) Tewkesbury, Gloucestershire. For the analysis, manufacturing and retail SMEs are categorised according to the level the floodwater reached in their respective premises (i.e. lightly, moderately, and severely) and then each category is investigated individually against each of the simulation experiments defined, which corresponds to different combinations of precautionary measures. As a result, this research has found that the flood resistance precautionary measures, which include a bund-wall being erected around SMEs’ premises, are effective for lightly and moderately flooded manufacturing and retail SMEs. However, the severely flooded manufacturing and retail SMEs simulated with a combination of flood resistance precautionary measures as well as business continuity/risk management tend to achieve higher production levels and service capability profiles, respectively.