This paper presents a new approach to the speed control of the T700 turboshaft engine using an adaptive controller of the self-tuning type for the power turbine. For the design of the SISO-STR, a SISO time-series model describing the relationship between the fuel flow as input and power turbine speed as output is used. The unknown parameters of this model can be updated recursively on-line by the RLS estimation algorithm. An adaptive self-tuning type controller are then designed by minimizing the expected value of a quadratic criterion which also includes an integrating effect to remove the steady-state speed errors. Three linearized models characterizing turboshaft engine without and with load dynamics are used to test robustness and performance specifications of the proposed self-tuning regulator in the digital simulation studies.