The objective of this study was to evaluate an improved bitterness sensor which has been developed to allow the precise and sensitive prediction of the bitterness of acidic bitter pharmaceutical active ingredients, using as examples nine non-steroidal anti-inflammatory drugs (NSAIDs). The bitterness of the nine NSAIDs was measured using a multichannel taste-sensing system incorporating a bitterness sensor, C00, which has a membrane surface with high hydrophobicity, and was developed to allow an enhanced hydrophobic interaction with acidic bitter substances. The sourness intensities of the nine NSAIDs were also determined in gustatory sensation testing and by a taste sensor using a sourness-sensitive membrane, CA0. The 'Change in membrane Potential caused by Adsorption' (CPA) of sodium diclofenac and etodolac were also determined in the presence of increasing concentrations of tartaric acid using membrane C00. Multiple regression analysis performed on the data on bitterness intensity obtained using the taste sensor and in gustatory sensation testing showed that CPA values from C00 could be used to predict the bitterness of the NSAIDs. The derived equation was y = -0.0413 × CPA + 0.3164, where y represents the predicted bitterness intensity. There were concentration-dependent changes in the bitterness intensities of diclofenac sodium and etodolac without any change in their sourness intensities. For diclofenac sodium and etodolac, there was a good correlation between predicted and actual bitterness intensities in the presence of increasing concentrations of tartaric acid. The taste sensor may be useful for predicting the bitterness intensity of acidic bitter pharmaceutical active ingredients such as NSAIDs.