Compound events occur when combinations of drivers and/or hazards contribute to a soci- etal/environmental impact (Zscheischler et al., 2020). Even if none of the individual drivers or hazards are extreme, their combination can produce extreme impacts. Assessing the potential for compound extreme events is therefore critical for effective risk management and mitigation planning. To determine the probability of compound events, statistical models are applied to time series data of the drivers or hazards, typically as the first step in the risk-analysis modeling chain. The MultiHazard R package is designed to enable practitioners to estimate the likelihood of compound events. Although the methods in the package are well-established in the scientific literature, they are not widely adopted by the engineering community despite guidelines increasingly mandating the estimation of compound event likelihoods. Functions within the package are designed to allow practitioners to apply their best judgement in making subjective choices. Inputs are time series representing the drivers/hazards; these may be historical observations or numerically generated with models (for the past or future). Outputs are: • Estimates of the joint return periods for specific combinations of drivers/hazards • Isolines with uncertainty bounds containing drivers/hazards with a specified joint return period along with the “most-likely” or an ensemble of events sampled along an isoline • Synthetic sets of events where the peak magnitude of at least one driver/hazard is extreme