In this paper, a novel classification method of two-dimensional poly-acrylamide gel electrophoresis images is presented. Such a method uses textural features obtained by means of a feature selection process for whose implementation we compare Genetic Algorithms and Particle Swarm Optimization. Then, the selected features, among which the most decisive and representative ones appear to be those related to the second order co-occurrence matrix, are used as inputs for a Support Vector Machine. The accuracy of the proposed method is around 94%, a statistically better performance than the classification based on the entire feature set. This classification step can be very useful for discarding over-segmented areas after a protein segmentation or identification process.