Provided is an electronic nose instrument based on gas sensitivity and gas chromatography and an online analysis method of multiple state parameters of fermentation and malodorous pollutant processes. The main constituent units of the instrument include a gas sensor array module, a capillary gas chromatographic column module, an automatic gas sampling module, and a computer control and analysis module. A single gas sampling period is T0=300-600s. Not only are two flow rates and two accumulative volumes of gas sampling unequal to each other, but also two starting time points are not synchronized to each other, between the gas sensor array module and the gas chromatography module. 3 pieces of sensitive information, i.e., a steady-state peak value, a corresponding peak time value and an area under a whole curve, are selected from a response curve of a single gas sensor with a 60s duration by the computer control and analysis module, or 48 pieces of gas sensitive information in total, and 21 pieces of sensitive information, i.e., 10 maximum peak values, 10 corresponding retention time values, and 1 area under the whole chromatographic curve, are selected from a semi-separation chromatogram with a duration T0−10 s. Furthermore, the cyclical online identification and intensity and quantitative estimation of multiple indices of odors for five fermentation or malodorous pollution processes with a maximum cyclical gas sampling period T=5T0 are realized by a modular deep convolutional neural network model according to a 69-dimensional normalized fused real-time sensitive pattern and an existing big odor data.