This paper addresses the problem of estimating the infrastructure to be made available for refueling alternative fuel vehicles as a function of the profitability thresholds required by the investment. A methodology has been devised based on sales forecasts for alternative fuel vehicles. These methods use discrete choice models in which the factor of refueling infrastructure, rather than being considered simply as one more attribute of the model, acts as a constraint on the choice set for vehicle buyers. This methodology is used to estimate the infrastructure of hydrogen refueling stations and electricity charging stations for Spain (8,112 population centers) in 2030. Evolution of fuel cell vehicles over the years 2016 and 2030 is also estimated and compared with forecasts for countries such as France, Germany and the United Kingdom. •With development of this optimization algorithm proposes a methodology to meet challenge posed by European Directive.•Methodology which corrects the effects of overestimating the AFV sales if infrastructure is considered as another attribute.•Methodology to locate new facilities depending on the protability parameter and its application to Spain Sensitivity analysis on the key parameter breakeven b, of the protability of the new facilities.•Compare the results obtained with the forecasts carried out for France, Germany, UK and Scandinavia for FCEV vehicles.