Аннотация:Dynamic optimization of crop photosynthetic productivity through continuous search for optimal environmental parameters can reduce the demand of board resources for space greenhouses. Drift of the crop environment can be tracked by an extreme regulator. We modeled search for etxremum of nonstationary wheat photosynthesis dependence on lighting intensity and air temperature with the use of gradient and modified simplex (Nelder – Mead) algorithms. It was shown that the least integral tracking error over the period of vegetation is achieved by a combination of the Nelder–Mead algorithm on the initial 20 steps and then, in the neighborhood of extremum trajectory, gradient algorithm. The Nelder – Mead algorithm was applied for a unit magnitude of the reflection coefficient, coefficient of triangles median stretch =1 and median contraction =2. In the gradient algorithm, gradient step 10-3 was chosen by the criterion of minimum integral racking error. These tracking algorithm parameters should be used in the system of adaptive optimization of crop illumination by light-emitting diodes in space greenhouse.Key words: space greenhouse, crop photosynthesis, dynamic optimization, search algorithm, extremum search errors.