This experiment was conducted to investigate the precision of empting-force feeding method for measuring the apparent metabolizable energy of corn starch diets in broilers. A single factorial completely randomized design was adapted in this experiment. Forty-eight 18-week-old yellowed-feathered broilers were divided into 2 groups with 6 replicates per group and 4 birds per replicate. The apparent metabolizable energy of corn starch diets was repeatedly determined at a force-feeding amount of 40 or 25 g using empting-force feeding method in 4 batches, respectively. The results showed as follows: 1) in the treatment of 40 g per bird for the force-feeding amount, there was no significant difference in the apparent metabolizable energy determined among 4 batches (P>0.05), and the variation coefficients of within-batch, between batch and total of apparent metabolizable energy were 5.39%, 1.89% and 5.27%, respectively; in the treatment of 25 g per bird for the force-feeding amount, there was significant difference in the apparent metabolizable energy determined among 4 batches (P<0.05), the apparent metabolizable energy of corn starch diets determined in the sixth batch was significant lower than that of the apparent metabolizable energy determined in the fifth and seventh batches (P<0.05), and the variation coefficients of within-batch, between batch and total of the apparent metabolizable energy were 1.97%, 1.53% and 2.36%, respectively. Compared with the treatment of 40 g per bird for the force-feeding amount, the variation coefficients of within-batch, between batch and total of the body weight losses, dry excreta from force-fed birds and gross energy output from force-fed birds of 25 g per bird for the force-feeding amount were lower. 2) In the treatment of 40 g per bird for the force-feeding amount, the contribution of the variance of dry excreta from force-fed birds, gross energy content of excreta from force-fed birds and force-feeding amount to the variance of apparent metabolizable energy of corn starch diets were 96.46%, 3.23% and 0.30%, respectively. In the treatment of 25 g per bird for the force-feeding amount, the contribution of the variance of dry excreta from force-fed birds, gross energy content of excreta from force-fed birds and force-feeding amount to the variance of apparent metabolizable energy of corn starch diets were 67.72%, 24.51% and 7.76%, respectively. In conclusion, the variation of apparent metabolizable energy of corn starch diets determined in the treatment of 40 g per bird for the force-feeding amount is greater than that in the treatment of 25 g per bird for the force-feeding amount. The variation of dry excreta from force-fed birds is the main source contributed to the variance of the apparent metabolizable energy of corn starch diets.