Bioimpedance properties of Fuji apple fruit with different freshness were investigated. Fruit with weight loss of 0, 5%, 10%, 15% and core-browning under room temperature were chosen as samples corresponding to descending freshness grades of 1,2,3,4,5, respectively. 126 observations of 14 bio-impedance parameters under nine frequencies of total 424 fruit from each grade were measured fruit by fruit with LCR impedance meter. Data were computed by using the sparse principal component analysis-linear classifier (SPCA-LDC) model. Results indicated that classification accuracy reached to maximum stably when the ratio of training sample to test sample of 9∶1 and former 39～45 principal component (PC) were adopted. The least 39 PCs were consisted of 44 from the 126 observations. Average rate of classification accuracy reached to 87.90% within 50 test repeats. 