This work addresses the problem o f signal resolution by partial least-squares regression (PLS) for differential pulse stripping voltammogranis. The performance o f PLS is demonstrated for an example of calibration with mixtures o f two analytes (TI+ and PbL+) with strongly overlapping voltammograms. The accuracy o f the determination o f concentrations for validation mixtures was found t o be considerably higher with PLS than with simple evaluation using currents observed at two potentials (requiring resolution of a system of two linear equations with two unknowns). Hence, I'LS is recommended for computer-assisted, on-line voltammetric analysis o f multicomponent systems.