Novel chemical sensors based on a time-dependent nonlinear response are reviewed. The strategy is to artificially mimic information transduction in living organisms. In taste and olfaction, information of chemical structure and concentration is transformed into nervous impulses in the nervous cell, i.e., time-dependent multi-dimensional information. Because the excitation and pulse generation in the nervous cell are typically nonlinear phenomena, it may be worthwhile to utilize the nonlinearity as the multi-dimensional information for molecular recognition. The principle of a "nonlinear" sensor is that a sinusoidal modulation is applied to a system, and the output signal is analyzed. The output signal of the sensor is characteristically deformed from the sinusoidal input depending on the chemical structure and concentration of the chemical stimuli. The characteristic nonlinear responses to chemical stimuli are discussed in relation to the kinetics of chemical compounds on the sensor surface. As a practical application, we introduced electrochemical sensors based on the differential capacitance, semiconductor gas sensors under the application of sinusoidal temperature or diffusion change, and a chemical sensor based on the spatio-temporal information. We demonstrated that mutli-dimensional information based on nonlinearity can provide quite useful information for the analysis of chemical species, even in the presence of another analyte or an interference with a single detector.