The inhibition of the aldose reductase enzyme (AR) is considered to be a promising approach to control chronic diabetes complications as well as a number of other pathological conditions. Thus considerable efforts are devoted to the development of aldose reductase inhibitors (ARIs) as possible pharmacotherapeutic agents. The establishment of adequate QSAR models would serve to this purpose. In the present study multivariate statistics was applied in order to analyse the AR inhibitory activity data of twenty three pyrrol-1-yl-acetic acid derivatives on the basis of essential molecular descriptors. The compounds contain one or two carbonyl keto groups, which serve as a bridge to link the pyrrole moiety to aromatic nuclei with or without further substitution. An adequate one component model with satisfactory statistics was obtained and validated for its robustness and predictive ability. The influence of the different descriptors in ARI activity is discussed. The derived model was further used to predict the activity of four independent compounds and the contribution of their specific structural characteristics in ARI activity was evaluated.