As passengers we would prefer to pay the cheapest fare available for our airline ticket. On the other hand airline companies wish to increase their revenue from its flown tickets. During the booking process of an airline flight, some passengers may arrive early to book their seats, others may decide to book just few days before departure or even on the day of departure. Airlines realise that they have to offer a variety of fares in order to differentiate between different types of passengers. Allocating seats to different fare classes for different types of passengers in such a way that would maximise the airline’s revenue requires yield/revenue management systems. There are two main steps in any revenue management system: Forecasting and Optimisation. Accurate prediction of passenger future demand for different fare classes improves the seat allocation recommendations resulting from the optimisation step. The work in this thesis concentrates on studying and analysing the behaviour of different passenger types towards different fare classes. We first formulate a Monte Carlo simulation model for the booking process. The model generates sample booking data for a flight on different fare classes by different types of passengers defined by the characteristics which affect their behaviour. Passenger behaviour is modelled using a customer utility function and a multinomial logit (logistic) model of demand. This sample booking data is then used in a Markov Chain Monte Carlo model in order to estimate the passenger choice parameters used in generating the booking data. These estimated parameters could be used then to classify any new booking data. The MCMC model uses the Metropolis Algorithm for its estimation process. We also examine briefly the computational feasibility of our approach using parallel processing.