This work makes a thorough study of the current operation of the service to disabled guests of the Barcelona - El Prat airport to make it more efficient. The most important problem they have to face nowadays is the uncertainty in the number of passengers that will need the service, and therefore is almost impossible to adapt the staff to the real demand. This derives to a malfunctioning of the service, which gives a bad image of the airport and maybe, economic penalties. The objective of this project, therefore, is be able to predict with the maximum possible accuracy how many incidentals there will be throughout the day, and more precisely, every hour, and hence be able to configure the workforce more efficiently. In order to be able to do this prediction, the airport has given to us the data corresponding to the service of the year 2013, with which we have made three studies. The first two studies were used to extract the data needed to do the subsequent simulation of the service. The first study is the calculation of the distribution of the probability that the data follows each day and hour. We have used the Pearson chi square test to see if our initial hypothesis could be accepted as valid or not. The second has been a study of the service times regarding the boarding gate. We have been able to demonstrate that the service time depends only on the boarding gate. To do so, we have divided the airport in zones, grouping the boarding gates with a similar average of service time and we have obtained that each zone follows a uniform distribution. The third study, the calculation of the waiting time of the passengers to be attended, is used to compare the quality of the service before and after applying our results from the simulations. These simulations have given to us that the number of PMR passengers is very similar to the real, so the previous statistic study is validated. Moreover, we have organized the service that the 90% or 95% of the passengers demanding the service will not wait more than the maximum waiting time established by the norm. This improves significantly the actual waiting times without the necessity of increasing the number of agents. However, the fact that we have studied all year globally by the lack of more data, our program oversize a little bit the service during the months when the demand is lower. This phenomenon could be improved in future studies by taking into account more data from subsequent years. These data could be obtained by doing studies of the growth of the airport during two or three years. Moreover, there are other possible improvements in order to do this study more accurate and, hence, optimizing better the available resources.