The discrete approximation of noisy data by algebraic polynomial with restriction of type equality is studied. The aimof the investigation is to obtain the fundamental properties of solution of the problem and development by them the new algorithm, more effective, in comparison with existing methods of the solution. The tasks of the research — gets the properties of the solution of the problem, presentation of the algorithm and the demonstration of its implementation. Research methodology continues P. L. Chebyshjov’s and Valle-Pussen’s method. Results. The criterion for optimality of the solution, which is a retrofit of the well-known in the theory of approximations of alternance P. L. Chebyshjov. Developed a rational algorithm, similar to the algorithm Vallee-Poussin. The conclusions. This problem has application to assess noise events at approximation to complex chaotic processes.