Noise removal is considered a primary and inevitable step for background correction in experimentally obtained Raman spectra. Employing an appropriate algorithm for a smoothing-free background correction technique not only increases the speed but also eliminates unwanted errors from the smoothing algorithms. Herein, we show a new smoothing-free method for background correction, which we developed by merging continuous wavelet transform and signal removal method, which in combination, could be applied to noisy signals without smoothing. We used wavelet transformation for suppressing the side effects of noise and eliminating peaks from the spectrum, thereby providing spectral sections purely related to the background to be used in the background correction process. We applied a range of statistical analyses to test the performance of this algorithm, wherein a low deviation in background correction procedure was observed. Additionally, when we tested this algorithm for experimentally obtained real Raman spectra, it showed good capability to correct background of noisy signals without the requirement of a smoothing process. Copyright © 2013 John Wiley & Sons, Ltd.