The invention relates to an electronic nose sampling data pre-calibration method based on a response curve differential characteristic. The method is suitable for adopting an electronic nose bionic sampling mode with a complete cycle in three cycles. The method comprises the steps that array sampling data acquired in an electronic nose bionic sampling manner are preprocessed in turn, and St is set as normalized relative conductivity change rate; St is differentiated through a sampling gradient method; the maximum of three ascending differential curves and the maximum of three descending differential curves of each sensor curve are extracted; the absolute value of three mean differentials of each sensor differential curve in a time period above a coordinate axis and the absolute value of three mean differentials in a time period above the coordinate axis are respectively calculated; the acquired maximum of the differential curve of each sensor is compared and judged; the acquired average differential of each sensor is compared and judged; and when the differential extremum and the mean differential meet the differential characteristic of bionic breath sampling, the sampling data of a sensor is accurately pre-calibrated.