A scaling model was built to calculate the activity of alpha emitting radionuclides in contaminated soil in the lysimeter field. Linear regression can be applied for the evaluation of radioactivity measurement data. Activities of the radionuclides 241Am, 238Pu, 239,240Pu and 90Sr obtained by experiments from real contaminated soils of the experimental lysimeter placed in a nuclear power plant (NPP) in Slovakia were evaluated using linear regression models with the method of least squares. A suitable scaling model for monitoring the 241Am, 238Pu, 239,240Pu alpha radionuclide activity was built using the regression triplet analysis and regression diagnostics. A regular designed scaling model opens the possibilities of longtime activity monitoring of these radionuclides, thus decreasing the number of necessary radiochemical analyses. The Fisher-Snedecor test, however, confirmed that the regression model for 90Sr activity monitoring by 241Am, 239,240Pu activity determination in contaminated soils can not be recommended.