The content of chlorides, nitrites, nitrates, phosphates and sulphates was used to classify 45 donkey's milk samples collected from different Italian regions. A method employing ion exchange chromatography with conductivity detector and chemical suppression was used. The quantitative results indicated phosphates (569.4-1304.4 mg kg(-1)) and chlorides (545.9-1757.9 mg kg(-1)) as being the most abundant anions, followed by sulphates (109.5-200.7 mg kg(-1)). The concentrations of nitrites and nitrates were found to be lower at 5.6 and 5.5 mg kg(-1) respectively. The data set was subdivided into three groups according to the region of origin of milk, and was statistically evaluated by analysis of variance (ANOVA). Concentrations of chlorides and nitrites showed a significant difference among farms (p < 0.001). In a first discriminant analysis procedure, functions based on linear combinations of the log(e)-transformed element concentrations of anions were generated to classify donkey's milk samples from different regions. In an alternative approach, a three-step discriminant analysis procedure to classify a milk sample was tested. The results obtained led to a correct classification of donkey's milk samples based on their anions content with 91-98% of the samples being correctly classified. The procedure proved to be very simple, so it could be used as an evaluation method for the traceability of donkey's milk, thus defending this unique product against fraud or commercial disputes.