For the development of highly active fungicides against Magnaporthe grisea, we studied the substituent effects at three sites (ortho, meta, or para) of R 1 and at various sites of R 2 aromatic rings in thiazoline derivatives at a 10ppm concentration for fungicidal activities against this target. Quantitative structural-activity relationships (QSAR) analysis to study the relationship between the substituent effects and the activities of the compounds was carried out by using multiple linear regression (MLR) and neural networks (NN). The results of the MLR and the NN showed good correlations (r 2 values of 0.849 and 0.884, respectively) between the selected descriptors and the activities in the training set. The descriptors, including a Hammett constant (σpR1), Connolly surface area (SAR2), and the substituent volume (SVRC2,31), play an important role for the activities of the compounds. Although the descriptors of an optimum MLR model were used in NN, the results were little improved by the NN, implying that the descriptors used in the MLR model included good linear relationships.