Control vector parametrization is one of the most frequently used techniques for determining numerically the optimal control profile for batch process optimization. In this article, we suggest the Karhunen-Loeve eigenfunctions as trial functions for obtaining a better representation of the exact optimal control profile while keeping the number of coefficients to be determined as small as possible. The procedure for finding the K-L trial functions is demonstrated on a batch electrochemical process for glyoxylic acid synthesis, and closed-loop optimization by control vector parametrization with K-L trial functions is studied by simulation. The results show that only two trial functions are needed for a satisfactory representation of the open-loop profiles, and as a consequence of the reduced solution space, the on-line computation need is greatly relieved, and the optimization algorithm converges more rapidly and more stably.