Cadmium (Cd), lead (Pb) as dangerous ecotoxicants pose a significant danger to agroecosystems and all living organisms. They enter the environment due to the action of anthropogenic-technogenic, military factors, due to global climate change, etc. Data analysis was carried out based on the results of a scientific and economic experiment conducted on dairy cows with different types of feeding in the Forest-Steppe soil-climatic zone of Ukraine. Cows were selected by the method of analogues for live weight, productivity with the third lactation period. The rations included feeds with an excess of Cd and Pb. The ecotoxicity of pollutants affected their transition from soil to plants (feed). The purpose of the research is to analyze the correlation relationship, construct regression equations between the concentration of heavy metals (HM) Cd, Pb in soil samples and feeds of different types of dairy cow diets, which will allow predicting the transfer of pollutants from soil to feed using environmentally safe milk production technologies, to carry out environmental monitoring and to assess environmental risks on cattle farms in a timely manner. Using the STATISTICA program, a correlation analysis was performed using the parametric Pearson correlation coefficient, taking into account the Kolmogorov-Smirnov and Lilliefors tests for normality, and the Shapiro-Wilk’s W test. To model the dependence between the variable U (concentration of heavy metals in feed diets) and the vector variable X (concentration of heavy metals in soils of agroecosystems), the construction of linear regression equations was used. The nature of the relationship was checked by constructing scatterplots; analysis of residuals ¬ for compliance with the law of normal distribution (Gaussian); assessment of the acceptability of the model as a whole ¬ by the level of probability using the ANOVA method; regression quality ¬ using the coefficient of determination R2. The interpretation of correlation data was carried out on the scale (Gorkavyi, 2019) r<0.3 - no connection, the variation of one variable almost does not explain the other, often has no practical significance; r=0.3-0.5 ¬ the connection is weak, weak correlation; r=0.5-0.7 ¬ the connection is average (moderate), a significant proportion of the variation is explained; r>0.7 ¬ the connection is close, strong correlation, high predictability of one variable by another. The analysis established in all cases a positive close relationship between the content of heavy metals in soil samples of agroecosystems and feed rations of different types of feeding r>0.7, which indicates a strong correlation dependence and high predictability of one variable for another. The correlation coefficient for cadmium concentration ranged from 0.799 to 0.971, for lead from 0.790 to 0.869 (p<0.05). The highest correlation was established for the content of Cd, where cows were fed a silage-hay-concentrate type ration r=0.971 (very high (strong) correlation) (p<0.05), for the content of Pb ¬ silage-hay ration r=0.869 (very high (strong) correlation) (p<0.05). The assessment of the acceptability of the model as a whole by the ANOVA parameter showed that the level of probability for the concentration of Cd is p=0.014, p=0.057, p=0.028, p=0.001, Pb ¬ p=0.051, p=0.025, p=0.032, p=0.062. The most acceptable model for the concentration of Cd in cows with silage-root crops, silage-haylage and silage-hay-concentrate type of feeding is p<0.05, for the content of Pb ¬ the model in animals with silage-hay and silage-haylage type of feeding is p<0.05. The corresponding linear regression equations y = 0.0645+1.5115×х (Cd), y = 0.277+0.861×х (Cd), y = 0.0283+1.4587×х (Cd), y = -11.1161+1.4595×х (Pb), y = 2.7024+1.1513×х (Cd), y = -19.5433+3.0059×х (Pb) were constructed to predict the transfer of pollutants from soil to feed in the diets of productive animals. The tested models give the most accurate result of predicting the concentration of Cd and Pb. A positive close relationship has been established between the concentration of Cd and Pb in the soils of agroecosystems and the diets of dairy cows with silage-root crops, silage-hay, silage-haylage, silage-hay-concentrate feeding types. The most acceptable models that will give an accurate result of predicting Cd in cows with silage-root crops, silage-haylage and silage-hay-concentrate feeding types; Pb ¬ silage-hay and silage-haylage. Further research ¬ conducting correlation and regression analysis between the content of heavy metals in feed and cow milk.