In vitro human iPSC-derived neuronal systems are an alternative platform for neurotoxicity testing to animal models and primary cultures. Microelectrode array (MEA), measurement system of the electrophysiological activity, are suitable to evaluate the neurotoxicity of compounds. We have previously reported the electrophysiological responses to known neurotoxicity compounds using MEA in human iPSC-derived neurons. However, the identification of analytical parameters to detecting toxicity of unknown compounds remains an important issue. We identify the analytical parameters enabling the separation of responses between neurotoxicity and negative control, and the separation the mechanism of action by using principal component analysis. By comparing the optimized analytical parameters of each testing compound with the standard deviation (SD) of negative control, we can predict general neurotoxicity in relatively quantitatively scale, for example, low risk for lower than SD range, medium risk for 2xSD, and high risk for over 2xSD. This predictability can help select appropriate concentration levels to avoid toxicity/adverse effects. In this study, we evaluated the toxicity risk of pesticides and industrial chemicals. We demonstrated that the toxicity risk can be detected by MEA measurement of human iPSC-derived neurons.