The frequency is an important operating parameter for protection, control, and stability of a power system. Due to the sudden change in generation and loads or faults in power system, the frequency is supposed to deviate from its nominal value. It is essential that the frequency be maintained very close to its nominal frequency. An accurate monitoring of the power frequency is essential to optimum operation and prevention for wide area blackout. As most conventional frequency estimation schemes are based on DFT filter, it has been pointed out that the gain error could cause the defects when the frequency is deviated from nominal value. This paper presents an advanced frequency estimation technique using gain compensation to enhance the DFT filter based technique. The proposed technique can reduce the gain error caused when the frequency deviates from nominal value. Simulation studies using both the data from EMTP-RV software and user defined arbitrary signals are performed to demonstrate the effectiveness of the proposed algorithm. The simulation results show that the proposed algorithm achieves good performance under both steady state tests and dynamic conditions.