Recent outbreaks of cryptosporidiosis raise serious concerns about the effectiveness of watershed management in controlling the risk of contamination by pathogens. A modeling strategy that takes into account the inherent randomness of the occurrence and transport of pathogens in surface water is important for accurate risk assessment and prediction of water contamination events. This paper presents a stochastic Markov model of microorganism transport, with distinct states of microorganism behavior capturing the microbial partitioning between solid and aqueous phases in runoff and soil surface, including the partitioning among soil particles of various sizes. A connection between the soil sediment and microbial transport is established through the incorporation of an erosion model (WEPP) into the microorganism transport model. Probability distribution functions of microorganism occurrence in time and space are constructed, and their properties are analyzed. A deterministic mathematical model of coupled pathogen and sediment transport is developed in parallel to parameterize the Markov process. The numerical values for the Markov model parameters were derived from two sets of experimental data. Model results show that areas with clay soils are more likely than sandy soils to contribute to contamination events and that the most influential transport parameters are the saturated hydraulic conductivity, rainfall intensity, and topographic slope.