To predict loin eye area in live pig using loin eye depth,traits of 286 carcasses and 357 live pigs were measured by standard method and B-mode ultrasound scanner,respectively. The traits data from the 286 carcasses (live weight at slaughter,back fat thickness,loin eye depth,loin eye width,loin eye area and lean percentage) and 357 live pigs (live weight at end,back fat thickness,loin eye depth and loin eye area) were analyzed to set up the prediction model of loin eye area,respectively and the prediction model was verified. The results showed that the correlation coefficient between loin eye depth and loin eye area of the carcass model (0.945) was higher than that of the living model (0.720); the determination coefficient (R2) of the carcass model (0.894) was higher than that of the living model (0.519). The predicted values were consistent with the measured values of the carcass model,while the predicted and measured values of the living model were more dispersed. The correlation coefficient between the predicted and measured loin eye area was 0.497 for carcass,and 0.672 for live pig,which were significantly positively related. There was no difference between the prediction and measurement data in the carcass model,while significant difference was observed in the living model. The results showed that it is reasonable to predict the live pig loin eye area with the carcass model,but further research is needed to improve the accuracy.