This paper proposes a new model reference adaptive system (MRAS) for sensorless control of a grid-connected brushless doubly-fed reluctance generator (BDFRG). The main BDFRG advantage over a conventional doubly-fed induction generator (DFIG) is the lack of slip-rings and brushes, hence the higher reliability, maintenance-free operation and lower cost. The desired active and reactive power variations are accomplished using a vector control scheme, which requires the rotor position and speed information in wind power applications with maximum wind energy extraction. Electro-mechanical shaft sensors, commonly used for this purpose, are prone to failures in harsh environments, like those encountered in off-shore wind turbines, and particularly with regular brush replacement intervals of DFIGs, making the BDFRG an attractive brushless alternative. The proposed controller design has been shown robust and capable of producing quality rotor angular position and velocity estimates by implementing an inverter-fed winding (secondary) current based MRAS. The realistic simulation results presented have proven the high accuracy and effectiveness of the estimation technique for different test scenarios of the BDFRG.