Keeping workers safe presents a continuing challenge in the agricultural industry. Risk assessment methodologies have been used widely to better understand systems and enhance decision making with a goal of reducing injuries and fatalities. This research applies probabilistic risk assessment to human safety in two agricultural production systems, taking into account uncertainties such as equipment variation, working schedules, and weather conditions. A comparative model was developed because it can be scaled up or down based on available data and allow inputs from categories defined broadly or specifically as necessary. In this model, risk is calculated by multiplying the probability of exposure to a hazard and the probability of injury, given that an exposure to the hazard has occurred. The probability of injury and exposure values are derived from the USDA Census and from the Survey and Bureau of Labor Statistics data from 12 states in the Midwest for each year from 1996 to 2011. The exposure and injury data were used to build probability distributions that were randomly sampled using a Monte Carlo simulation. The output of the simulation demonstrates that corn has a higher risk of worker injury than biofuel switchgrass over a ten year period in the Midwest. A Monte Carlo simulation and a sensitivity analysis were run to determine the greatest contributing factors to worker injury risk within each production system. Harvest operations in both corn and biofuel switchgrass production systems were determined to be the greatest contributing factor to worker injury risk.