This paper presents a compression algorithm for power quality (PQ) signals, which does not separate the signal into stationary and transient components. The proposed algorithm is composed of a number of different compression techniques which are combined to improve both compression ratio and total computing time. This method is based on a sequence of three phases. The first phase transforms the signal to the frequency domain using the FFT algorithm. In a second phase the signal is approximated by a polynomial approximation algorithm, which is finally compressed by a lossless compression algorithm. The proposed method is characterized by high compression ratio and low computational cost, making its implementation suitable for embedded systems.