This paper is concerned with the investigation of an optimal form of excitation for external control of long-term ventilation of the lungs. An up-to-date model of the respiratory system is considered and the minimization of work done by the lungs during the respiratory period is chosen as the performance index for determination of the optimal excitation. With this optimal excitation, whose frequency is determined in terms of the natural resonant frequency of the lungs, an optimal adaptive controller is synthesized adopting the Bellman dynamic programming technique. The adaptive controller minimizes the harmful effects arising due to the change of respiratory parameters for controlled ventilution.