In this paper, the spatial distribution of fluoride (F -) and manganese (Mn) in groundwater was described and the overall groundwater quality was assessed with the entropy weight water quality index (EWQI) in Weining Plain, northwest China. The probabilistic health risk of Fand Mn in groundwater was assessed by the Monte Carlo stochastic simulation method. The results show that 50 groundwater samples (a total of 144) belong to poor and very poor quality because of agricultural activities, industrial development, and the local hydrogeological conditions. Groundwater quality in the upper reaches of the Yellow River is better than that in the lower reaches. The average health risks faced by children and adults are 4.59 and 0.62, respectively. If the uncertainty of the model parameters are considered, the mean risks faced by children and adults are 6.84 and 0.92, respectively, indicating that the health risk to residents, especially children, cannot be ignored. Compared with dermal contact, drinking water intake is the main exposure way to harm residents' health. Fcontributed higher to the health risk than Mn, and children face greater risk than adults. Stochastic simulation can reflect health risks more comprehensively than the deterministic methods. The study results are helpful for decision makers to take measures for the safe supply of drinking water in the study area.