Aldose Reductase (AR), the key enzyme of the polyol pathway catalyzes the reduction of glucose to sorbitol using nicotinamide adenine dinucleotide phosphate as an essential cofactor, has been demonstrated to play an important role in the pathogenesis of diabetic complications. Self Organizing Molecular Field Analysis (SOMFA), a novel three-dimensional quantitative structure activity relationship (3D-QSAR) method has been used in present case to study the correlation between the molecular properties and the aldose reductase inhibitory activities on a series of 5-arylidine-2, 4-thiazolidinedione. SOMFA calculations for both shape and electrostatic potentials were carried out. The master grid maps derived from the best model has been used to display the contribution of electrostatic potential and shape molecular field. The statistical results showed good cross-validated r(2)(CV), non cross-validated r(2), F-test set and significant predictive ability indicated by r(2)(pred). All analysis of SOMFA model may provide useful information in the design of new aldose reductase inhibitors with improved spectrum of activity for management of diabetic complications.