A stochastic concept for highway capacity analysis is presented. The capacity of a highway facility is regarded as a random variable instead of a constant value. Thus, the stochastic approach provides new measures of traffic flow performance based on aspects of traffic reliability. A method for the estimation of capacity distribution functions from empirical data based on statistical methods for lifetime data analysis is introduced. This method is derived for the analysis of freeway capacity. However, the stochastic approach also is shown to be applicable to intersections. Results of the analysis of data samples from freeway sections in Germany indicate that freeway capacity is Weibull-distributed with a considerable variance. A Monte Carlo technique based on the stochastic description of capacity is proposed to quantify freeway traffic performance over a whole year. This technique also provides a quantitative assessment for oversaturated conditions.