The topic of this paper is methodology for improvement of the efficiency of the downhole oil production control algorithms. The approach is based on the use of the well-known multi-cascade principle, in which the cascades represent the control algorithms of separate local layers of the process. The results of investigation of the multi-layered control principles on the example of the downhole oil production control process show the efficiency of application of the cascade control principle. The downhole oil production is considered as a multilayer control process: 1) Input layer - the local process of creation of pressure in an electrical submersible pump (ESP); 2) Inner layer - the local process of the fluid motion in the pump-compressor pipes; 3) Output layer - the local process of the fluid motion in the casing well. According to the principle of the cascade control, the control algorithm consists of several cascades (local control algorithms). Each local algorithm calculates the set value for the next local algorithm on the basis of the current value of the output parameter for the respective local process. As a result, the control algorithm immediately takes into account the current disturbances of all the previous layers. Duration of the compensation process of the local layer disturbance is determined by its own time constant. The paper presents a methodology for the local control algorithms design. The design procedure begins from the first (input) process layer. The design of the local algorithms is determined by the criterion of the minimum regulation time with the use of the model of the local process layer. The efficiency was assessed from the point of view of the influence of the inner layers' disturbances on the output parameters of the considered technological process. Analysis of the results of the computation experiments demonstrates a significant decline of the influence of the inner disturbances on the output parameters, when the multi-cascade control algorithm is compared with the use of the conventional control algorithms.