Absorbing controllable Markov processes as models of some technical and/or biological objects behavior at their lifetime are proposed. A problem of optimal control of these objects during their lifetime for two-level control policies with respect to two objective functionals as control goals are considered; these goals are: the maximization of mean lifetime and the mean reward during it. The closed formulas for the main characteristic of the process and objective functional are given and some numerical examples of optimal policies calculation for both functionals are done.