We report on an application of a simultaneous perturbation stochastic approximation (SPSA) algorithm to filtering systematic noise (SN) with nonzero mean value in photoemission data. In our analysis, we have used a series of 50 single-scan photoemission spectra of W (110) surface where different SNs were added. It was found that the SPSA-evaluated spectrum is in good agreement with the spectrum measured without SN. On the basis of our results, a wide application of SPSA algorithm for evaluation of experimental data is anticipated. (C) 2008 American Institute of Physics.