The applicability of Fujita's inorganic (i) and organic (o) characters as descriptors for predicting bioconcentration factor (BCF) in fish was investigated with a wide variety of organic chemicals. Among 612 BCF data recently released by National Institute of Technology and Evaluation (NITE), the values for the chemicals with a molecular weight of less than 600 and the 1-octanol/water partition coefficient (log P) of less than six were extracted and analyzed. By applying theoretically derived model equation, a good relationship between these BCF and each Fujita's sigma i: sigma o was established. Statistical analyses and model validations revealed that the estimations of our approach were very excellent. More precise predictions were attained than those by using other published models, especially, for chemicals such as disperse dye having hetero atoms.