The paper includes a stochastic model of the optimal manpower planning in dynamic organizational systems based on the renewal theory of employee turnover. We come up with a multi-level hierarchical staffing model, which involves random departure times, replacement processes based on skills and cost optimization in the presence of uncertainty. The offered methodology expands classical renewal theory by adding state-dependent renewal functions and multi- dimensional counting processes to the dynamics of the contemporary workforce. The theory we have developed provides closed-form solutions to optimal staffing levels when subjected to a number of organization constraints and Monte Carlo simulations confirm the strength of the model when it comes to a number of industries. The empirical validation based on the data of three multinational corporations shows that the workforce stability (15-23% decrease in the understaffing cases) and cost efficiency (12-18% decrease in recruitment costs) has improved significantly as opposed to the traditional deterministic methods. The practical applicability of the model is further facilitated by the fact that development of adaptive algorithms is done to suit real-time organizational changes and external market dynamics.