In design of adaptive orthogonal frequency division multiplexing (OFDM) transmission, Signal-to-Noise Ratio (SNR) is well-known to be an informative performance measures. Channel impairments, specifically in wireless communication degrade the performance of accurate SNR measurement and said problem becomes more serious in low SNR regime. As a result it reduces the performance of adaptive transmission in OFDM systems. To alleviate this problem we develop a pilot based Pre FFT (time domain) Signal-to-Noise Ratio (SNR) estimator in the presence of frequency selective Rayleigh fading channel. A novel time domain SNR estimation technique is proposed that accurately measures the Signal-to-Noise Ratio in low SNR regime, where signal plus noise power is evaluated by using autocorrelation of received signal and signal power is indirectly estimated from pilot power using cross correlation. Simulation results show that proposed method has very less bias and very close to the actual SNR values. © Springer International Publishing Switzerland 2015.