A computer spreadsheet model of paratuberculosis in dairy herds was developed using Reed-Frost methods to calculate the number of new infections, and Markov chain methods to calculate culling probabilities in each time period. The model is dynamic; annual paratuberculosis prevalence rates change as infected animals join the herd or are culled. Output from the model is illustrated by graphing disease prevalence against time. Seven variables are specified at the initial stage of the model: herd size, annual herd birth rate, annual herd replacement rate, number of infected cows at time zero, number of herd replacements purchased each year, risk of purchasing a M. paratuberculosis-infected heifer, and number of effective cow-calf contacts per year. Sensitivity analysis of five of these variables was performed. All variables affect the course of paratuberculosis spread in herds, but the model is most sensitive to the effective contact rate. This is consistent with the findings of other infectious disease models and recommendations on paratuberculosis control: minimize cow-calf contact to prevent transmission of the disease. The model is difficult to validate, however, because data on paratuberculosis prevalence changes in dairy herds in the absence of control measures is not available. A full description of the methods and sensitivity analysis on the model variables will appear in an upcoming issue of the Journal of Preventive Veterinary Medicine.