In this paper, a novel frequency estimation algorithm based on discrete Fourier transform (DFT) and second derivative is proposed. The proposed algorithm consists of three stages, namely, DFT, second derivative, and frequency estimation. First, the input signal is decomposed using DFT into two orthogonal components, the real and imaginary parts. In other words, the signal is filtered with cosine and sine filters. Secondly, the second derivatives of the two orthogonal components are obtained by numerical approximation. Because this step can cause an error, the central difference approximation for five-point second derivative is used for error reduction. Finally, the two orthogonal components and their second derivatives are combined to estimate the power frequency considering a zero-crossing problem and the gains of finite impulse response (FIR) filters. The performance of the proposed algorithm is evaluated considering frequency changes when generating the test signals. Besides, a dynamic condition which is simulated in a 5-bus transmission system modelled by PSCAD/EMTDC is also considered. The performance evaluations show that the proposed algorithm is an effective method.