Due to environmental pollution, heavy metals such as Pb, As, and Cd may contaminate raw milk and can involve serious systemic health problems if they are consumed in excessive concentrations. This study investigated the spatial distribution of Pb, As, and Cd in raw milk produced in the 10 main milk producing areas in China. The contents of Pb, As, and Cd in 996 raw milk samples [i.e., 100 milk samples per area except for 2 area (n = 97, n = 99)] were measured by ICP-MS after microwave-assisted acid digestion. Non-parametric Kruskal-Wallis test were performed to study the differences of Pb, As, and Cd between areas. Spearman correlations were calculated to assess the relationships between the studied heavy metals. Then, the spatial distribution of Pb, As, Cd was studied by ordinary kriging estimates within the studied areas. Cross-validation was used to assess the robustness of the distribution map. Mean values of Pb, As, and Cd were 1.75, 0.31 and 0.06 μg/L of milk, respectively. Levels of Pb in 1.20% (12/996) of collected samples were above the maximum residue limit (MRL) imposed by the European Union (0.02 mg/kg). All samples were below the Chinese MRL (i.e., 0.05 mg/kg for Pb, 0.1 mg/kg for As). High coefficient of variation were obtained within area suggesting a large variability of those metal contents in milk within regions. This shows the need to conduct a reflection about the best way to collect samples if this kind of pollution in milk want to be studied on a long period. Pb-Cd, As-Cd, Pb-As showed positive significant correlations in 9, 6, and 5 areas, respectively. Correlation values ranged between 0.20 and 0.60. However, these correlations changed between areas suggesting different pollution origins. Based on the ordinary kriging estimates, Pb, As, and Cd showed different spatial patterns following the studied area. Based on the cross-validation, the root mean square error was not closed to the average standard error in some areas. This leads potentially to wrong predictions. The high density of sample collection may lead to this result. Further studies could implement a more appropriate sample collection to clarify the relationships between the contamination of raw milk by heavy metals and the herd environment.