This thesis deals with the issue of travel time reliability. The study of travel time reliability emerges from the fact that in many situations, travel time is random. Many events can change the travel time forecasted by operators or expected by users. Moreover, a tradeoff may exist between time and reliability benefits when evaluating socio economic appraisal of transport infrastructure. However, since reliability is still difficult to integrate in this type of evaluation, investment projects’ collective profitability is underestimated and often postponed. Thus, three main issues of travel time reliability analysis emerge: measurement, monetary valuation and implication for cost benefit analysis. This thesis is organized in three chapters. The first chapter adapts the measure of travel time reliability typically used in the road transport context to the collective modes (rail and air, in particular). We also develop a new reliability measure: the Delay-at-Risk (DaR). DaR is an implementation of the Value-at-Risk (V aR) measure into the transport economic framework. The DaR seem to be relevant and understandable information for the users, especially to plan their travel and avoid missing their connections. The main objective of the second chapter is to define the users’ willingness to pay to improve travel time reliability. We present a theoretical framework based on decision theory under risk. We introduce the concept of reliability-proneness (i.e. travel time risk aversion) and prudence. We develop new measures of travel time reliability expressed as risk premium: the reliability-premium and V OR. The reliability-premium is the maximum amount of additional travel time that an individual is willing to accept to escape all the risk of travel time. The V OR is defined as the maximum monetary amount that an individual is willing to pay to escape all the risk of travel time. Furthermore, we also establish the link with attitudes towards risks of travel time (aversion and prudence) and the impact of the value of travel time (V TTS) and the value of reliability (V OR). The final chapter of this thesis integrates reliability in investments project’s socioeconomic appraisal. More particularly, it allows to determine users’ surplus valuation. We highlight a diffusion effect of reliability benefits with regard to travel time benefits. Thus, we propose recommendations regarding the tradeoff between projects that generate time benefits compared with reliability benefits, according to the monetary values of travel time (V TTS) and reliability (V OR).