Gas-to-gas membrane humidifiers are the preferred option for humidification in automotive fuel cell systems, because they are compact, induce low parasitic losses and require no moving parts or control mechanisms. Understanding the performance of the humidifier as a function of its operating conditions is crucial for the efficiency of the fuel cell system. Therefore zepp.solutions BV, a system integrator of hydrogen fuel cell systems for the clean and mobile power generation, demanded advanced insight into the performance of their humidifier. Two models were developed to describe the heat and mass transfer in a gas-to-gas shell-and-tube membrane humidifier, furthermore the performance of the humidifier was extensively determined via experiments. The novelty in the model lies in the comprehensive combination of advanced correlations describing mass and heat transfer in the humidifier. Experiments at various mass flows, temperatures and pressures were performed on a Fumatech H20N humidifier, a gas bubbler has been designed and developed to humidify the gas stream. Heat transfer was measured to increase linearly with inlet temperature difference and was limited by convective heat transfer. Water transfer was found to be limited by diffusion and to increase exponentially with wet inlet temperature, dry air mass flow only had a small impact. The impact of mass transfer on heat transfer can not be neglected, due to latent heat transfer and transfer of energy by water being transported. For the analytical and numerical model, the heat and mass transfer coefficients are based on correlations found in literature. The membrane diffusion coefficient and both convective mass transfer coefficients vary over an order of magnitude or more in literature. The numerical model can predict the latent effectiveness with acceptable accuracy, the sensible effectiveness is mainly overestimated. For eighty percent of the cases the predictions are within (+.0061 to .22) and (-.11 to +.056) for the sensible and latent effectiveness, respectively. A further improvement requires more information about the membrane diffusion coefficient of the specific membrane material and appropriate correlations for the shell side Sherwood numbers taking tube placement irregularity into account.