In the case of a fire disaster in a uni-directional road tunnel, it is important to determine the critical ventilation velocity to prevent the backflow travelling toward the tunnel exit where vehicles are stopped. The critical ventilation velocity is horizontal velocity to prevent hot smoke from moving toward the tunnel exit. According to Froude modelling, the model tunnel whcih was 300mm in diameter and 21 m in length was made of acryl tubes. Inner section of acryl tubes was clothed with polycarbonate. 1/20 scaled model vehicles were installed to simulate the situation that vehicles are stopped in the tunnel exit. Methanol in a pool type burner was burned in the middle of tunnel to simulate a fire hazard. In this study, the basis of determining the critical ventilation velocity is the ventilation flow rate that is able to maintain the allowable CO concentration in the tunnel section. We assumed that the allowable CO concentration was backflow dispersion index. Futhermore, We intended to find out CO distribution and temperature distribution according as we changed ventilation velocity. The results of this study were that no backflow happened when ventilation velocity was 0.52 m/s in the case of 5.75 kW. If we adapt these results of a fire disaster releasing 10MW heat capacity in real tunnel which is 400m in length, no backflow happens when ventilation velocity is 2.31m/s. After we figured out dimensionless heat release rate and dimensionless ventilation velocity of model test and those of real test to verify experimental correctness, we tried to find out correlation between experimental results of model tunnel and those of real tunnel.