At present, the main way of strawberry planting in our country is facility cultivation. Due to soil continuous cropping, improper temperature and humidity control and other problems, it is easy to cause diseases and pests, and the use of drugs by farmers has increased significantly, and the problem of pesticide residues in strawberries is more prominent. In this paper, an electronic nose system for detecting pesticide residues on strawberries was designed. The electronic nose system was improved to address the problem of poor recognition of the electronic nose due to the low concentration of the volatile odor of the residual pesticide and its susceptibility to the odor of the strawberry itself. Structurally, a bionic air chamber is designed by drawing on the structural characteristics of the human nasal cavity. The structure of the bionic gas chamber was optimized using computational fluid dynamics (CFD) simulations to ensure the quality of the signal at the acquisition end. Algorithmically, a classification model (NRBO-BP) based on the trap-avoidance operator (TAO) improved newton-raphson search rule (NRSR) optimizing back-propagation BP neural network was established to improve the classification algorithm's effect on the recognition of low-concentration signals. Strawberry containing carbendazim and imidacloprid and their mixed pesticides were detected by electronic nose. The results showed that the NRBO-BP classification model based on the bionic air chamber electronic nose had an accuracy of 93.44% and a recall of 94.16%. NRBO-BP classification model was generally higher than the 88.33% of the PSO-BP model and the 83.33% of the BP neural network, and was able to accurately detect pesticide residues on strawberries. It can be used as a rapid method for the evaluation of strawberry quality and safety.