The electronic tongue system based on virtual instrument technology was used to qualitatively analyze the aged wheat with four storage years to achieve rapid and objective evaluation and analysis of aged wheat with different storage years. In view of the complex output signal of the electronic tongue and the large amount of data, the wavelet packet transform was used to extract the eigenvalues of the original data to reduce the data dimension and reduce the data size. On this basis, the improved fish swarm algorithm was used to optimize the parameters of the extreme learning machine, and the analysis model of wheat storage age was established. The model was used to qualitatively analyze the aged wheat with five storage years. The experimental results showed that the model had better classification effect and the classification of WPT-IAF-ELM was compared with genetic algorithm and particle swarm optimization ELM algorithm respectively. The effect was better, and the training set correct rate, test set correct rate, overall classification accuracy and Kappa coefficient were respectively 96%, 92%, 95% and 0.91, which indicated that the proposed combined model had better classification effect for food.