Foam lift is one of the most cost effective methodologies for unloading gas wells. The surfactants are either injected intermittently or continuously to lift the liquid to the surface. By reducing the gravitational gradient and increasing the frictional gradient, the critical velocity at which liquid loading occurs is shifted to lower gas velocities. Currently, we do not have a methodology to predict the critical velocity (at the transition boundary of annular and intermittent flow) and the pressure drop under foam flow conditions. To address this, we measured several foam flow characteristics in both small scale and large scale facilities. Small scale facility involved measurement of foam carryover capacity as a function of time and surfactant concentration. Large scale facility involved measurement of liquid holdup, pressure drop, fraction of gas trapped in foam and foam holdup in 40-ft 2-in. and 4-in. tubing. We developed closure relationships for liquid hold up, foam holdup, fraction of gas trapped in the foam and interfacial friction factor by combining the small scale data with the data collected in the large scale experiments. These closure relationships are applicable to four different surfactants tested. A new transition criterion was developed and successfully used to predict onset of liquid loading under foam flow. Using a force balance over the gas core in annular flow, we developed a new procedure to calculate the pressure drop under foam flow conditions. We compared our model results with actual measurements in the large scale facility. Our model was reasonably able to predict the pressure drop within +/- 30%. The reason for such a large variance is that the small scale facility was not able to capture all the characteristics of the foam which were observed in the large scale facility. It is very difficult to reproduce the foam characteristics exactly in two different experiments. This is discussed further in this paper. The procedure developed is the only one currently available to calculate the pressure drop under the foam flow conditions using the small scale data. It is superior to conventional annular flow pressure drop prediction models which are currently available in the literature. (C) 2016 Elsevier Ltd. All rights reserved.