The paper describes a model which allows studying optimization problem for the partial privatization of a state-owned enterprise. The model assumes the possibility to choose the moment of privatization, as well as the stochastic behavior of the enterprise’s profits and its change after privatization. The proposed scheme allows to find the optimal (in NPV criterion of the buyer) time of privatization and the optimal (in budgetary effect criterion) degree of privatization. In the case when the profit is modeled by random processes of geometric Brownian motion, we study the structure explicit of the optimal degree of privatization. For the case of linear “privatization multiplier” it is obtained the complete characterization of the domains (in the space of model parameters) with “no privatization”, “partial privatization” and “full privatization”. We derive formulae for the optimal degree of privatization, and study the its dependencies on the profit parameters (average growth rate and volatility) and tax burden. In particular, it is shown that optimal degree of privatization falls when volatility of the profit grows.