The signal-to-noise ratio (SNR), the standard measure of analog noise attribute, is used to assess the quality of the final demodulation signals in a orthogonal frequency division multiplexing (OFDM) system. The approach is based on the statistical properties of a random process after the Fourier transform. A simple form of least-square SNR estimator in the frequency domain is presented, which drastically reduces the computational complexity since no timing alignment is necessary. The SNR estimation is used to evaluate not only the adverse impairment of the interference in an external channel, but also the severity of different distortions caused by non-ideal operations of the internal components. It is shown that with an effective SNR estimation, quantified measurement can be achieved in a short simulation time for such distortions as caused by non-ideal channel estimation, signal clipping, carrier frequency offset, and many other sources in the system. This leads to more rational trade-offs in the hardware design to optimize the performance of the system under given implementation constraints.