Abstract: We detected tilapiaat different frozen storage time by an electronic nose (PEN3) and analyzed the effect of frozen storage time on its flavor. The data was analyzed by PCA and LDA methods. The results show that electronic nose can discriminate tilapiaat different frozen storage time especially well by PCA and LDA methods. The first and second main axes generally represent the main characteristic information of frozen tilapia, whose sum of variances is 91.69% by PCA method. The sum of variances of the first and second main axes is 92.39% by LDA method which has good concentration and single tendency. In addition, sensors W5S, W1S, W2S and W1W in electronic nose (PEN3) analyzed by LOADINGS method play a key role in the detection. Therefore, the screen matrix arrangement of sensors of electronic nose can be optimized according to specific experimental conditions.