The composition of Devonian oil has been simulated using an experimental true-boiling-point (TBP) curve, and the physicochemical properties (density, molecular weight, enthalpy, etc.) of the feedstock arriving at the plant have been calculated. The optimization criterion proposed is the calculation of the yield of light fractions. The optimization problem with constraints on variables (initial and final boiling points of light fractions) and constraints in the form of inequalities (overlapping neighboring fractions) using the function fmincon has been solved and implemented in the programming environment Matlab. A nonlinear dynamic model of the oil rectification process has been developed and implemented. The Matlab ode 15s solver and the explicit second-order Euler method have been used to develop the model. The model in the form of a system of algebraic-differential equations has been applied to the distillation column control system of an atmospheric pipe still at a small oil refinery. In the case of controlling according to the model, the column profile temperatures are maintained at optimal (preset) values and the quality of the petroleum products is also maintained. A comparison has been made between two operating modes of the distillation column in the processing unit. The results of the study suggest the following. The method of controlling in accordance with the model provides an increase in the recovery of light products through the use of optimization and the dynamic model, shortening of the overlap intervals of the temperatures of the gasoline and diesel fractions, and a decrease in deviations of the boiling points of petroleum products.