Wildland fire scientists and land managers working in fire-prone areas require spatial estimates of wildfire potential. To fulfill this need, a simulation-modelling approach was developed whereby multiple individual wildfires are modelled in an iterative fashion across a landscape to obtain location-based measures of fire likelihood and fire behaviour (e.g. fire intensity, biomass consumption). This method, termed burn probability (BP) modelling, takes advantage of fire spread algorithms created for operational uses and the proliferation of available data representing wildfire patterns, fuels and weather. This review describes this approach and provides an overview of its applications in wildland fire research, risk analysis and land management. We broadly classify the application of BP models as (1) direct examination, (2) neighbourhood processes, (3) fire hazard and risk and (4) integration with secondary models. Direct examination analyses are those that require no further processing of model outputs; they range from a simple visual examination of outputs to an assessment of alternate states (i.e. scenarios). Neighbourhood process analyses examine patterns of fire ignitions and subsequent spread across land designations. Fire hazard combines fire probability and a quantitative assessment of fire behaviour, whereas risk is the product of fire likelihood and potential impacts of wildfire. The integration with secondary models represents situations where BP model outputs are integrated into, or used in conjunction with, other models or modelling platforms.