Potassium-magnesium ores of the Chelkar deposit in Kazakhstan have a complex mineralogical composition and contain soluble sodium salts, which are an impurity component in the processing of ore into potassium fertilizers. In this regard, enrichment of natural potassium salt by incomplete dissolution was studied. Based on the experimental results the optimal washing mode was determined which ensures maximum removal of sodium salts from the natural salt composition and minimal losses of potassium salts with washing water. Mathematical models of the process under study were obtained through statistical processing of the experimental data. The results have shown that statistically, reliable regression equations are ones that describe potassium and sodium content in the solid phase relative to the process duration and water consumption for washing. The models can be used to operate the potassium ore enrichment process and to predict dependencies under various process conditions. 3D visual analysis of the process using regression surfaces for all significant equations has revealed the relationship between potassium ore enrichment indicators and process parameters.