To examine the effects of physical training on glucose effectiveness, insulin sensitivity and endogenous glucose production (EGP) in middle-aged men, stable-labeled frequently sampled intravenous glucose tolerance tests (FSIGT) were performed on eleven exercise-trained middle-aged men and twelve age-matched sedentary men. The time course of EGP during the glucose tolerance test was estimated by the nonparametric stochastic deconvolution. Glucose uptake-specific indices of glucose effectiveness (S_G²* x 10²: 0.81±0.08 vs 0.60±0.05, min^-1.d^.kg^-1, P<0.05) and insulin sensitivity [S_I²* x 10⁴: 24.59±2.98 vs 11.89±2.36, min^-1.(µU/ml)^-1.dl^.kg^-1, P<0.01], which were analyzed using the two-compartment minimal model, were significantly greater in the trained group than in the sedentary group. Plasma clearance rate (PCR) of glucose was consistently greater in the trained men than in sedentary men throughout FSIGT. Compared with sedentary controls, EGP of trained middle-aged men were higher before glucose load. The EGP of the both groups was similarly suppressed by ~70% within 10 min, followed by an additional suppression after insulin infusion. The EGP returned to the basal level at approximately 60 min in the trained men and at 100 min in the controls followed by its overshoot, which was significantly greater in the trained men than in the controls. In addition, the basal EGP was positively correlated with S_G²*. The higher basal EGP and greater EGP-overshoot in trained middle-aged men appears to compensate for the increased insulin-independent (S_G²*) and -dependent (S_I²*) glucose uptake in order to maintain glucose homeostasis.