Rowers transfer power output (PO) to the handle with the movement of their body toward back and forth which is defined as internal work, however, accumulated oxygen deficit (AOD) is estimated from only PO. This study aimed to investigate the validity of a newly proposed method, which estimates oxygen demand from PO and internal work, by comparing the relationships between AOD and blood lactate related parameters. Eight male university rowers performed a 2-minute supramaximal test, and blood lactate samples were obtained for 90-min of the following recovery period to evaluate blood lactate concentration at the end of the exercise ([La]b(0)), its peak value ([La]b peak), and a quantity of lactate accumulation (QLaA). Higher AOD was observed in the new method relative to the conventional method (change of mean, 13.8 %; 90 % confidence limits (CL), ± 5.1 %). Furthermore, the AODs estimated by the both conventional and the new methods were likely to correlate with QLaA (r = 0.65; 90 % CL, ± 0.43 and r = 0.59; 90 % CL, ± 0.47, respectively), however, unclear to correlate with [La]b(0) (r = 0.12; 90 % CL, ± 0.62 and r = -0.03; 90 % CL, ± 0.63, respectively) and [La]b peak (r = 0.13; 90 % CL, ± 0.62 and r = 0.10; 90 % CL, ± 0.62, respectively). In conclusion, our results suggest that the both of estimation methods provide AOD being likely to show a moderate correlation with QLaA, however, the value is higher in the new method.