A probabilistic risk assessment (PRA) using two-dimensional Monte Carlo analysis was followed to inspect the health risk associated with consumption of groundwater contaminated with nitrate in 26 wells located in rural areas of Malayer, Iran. In this technique, probability distributions were assigned to the concentration of nitrate in groundwater, daily intake rate of water, frequency of exposure together with total duration of exposure and the risk levels were worked out for adults and children, accordingly. In addition, four scenarios were investigated with an emphasis on the effect of correlation between exposure frequency and ingestion rate of water on estimated risk values. It was indicated that inclusion of correlation between parameters would influence the upper median quartile values of estimated risk however the total impact on the results of health risk assessment is not significant. Moreover, considering the 3rd quartile value of risk level for the total size of the study area, the risk level related to children was higher than that of the adults with respective values of 0.621 and 0.989. The spatial variation of nitrate was considered using ordinary kriging and a novel combined method of random forest and spatial proximity as covariate. It was concluded that the predictions made by ordinary kriging were more accurate than the random forest technique which can be attributed to the small data sets used in this research.