Through several contractions, stiff competition, and increasing passenger expectations, airports must evolve continually. One of the main avenues for this has been improving the efficiency of the security check- points, which are airports’ primary bottlenecks. Operational optimisation methods, such as resource and task scheduling are relatively mature fields of research, however, they require accurate forecasts. Current forecasting approaches seldom use useful information such as the flight schedule, nor are they able to re- present uncertainty or integrate real-time information. Therefore this paper aims to develop and evaluate a real-time probabilistic security checkpoint arrival rate forecasting model by utilising a Bayesian frame- work. This is achieved using a probabilistic programming language to create a bottom-up model, where per-flight arrivals are predicted. The passenger arrival rate for each flight is determined by estimating the total number of passengers and their temporal arrival distributions probabilistically. The combination of arrival rates from all flights in the flight schedule then provides the full checkpoint forecast. Furthermore, an updating scheme is proposed, that updates the expected number of passengers for each flight through Bayesian inference. Results show that the static forecasting model has promising performance, while suc- cessfully capturing uncertainty. However, the proposed real-time updating approach does not function as intended, due to a consistent negative bias. This has been attributed to a fundamental asymmetry present in the problem. Finally, this study includes an application of lane requirement estimation, which yielded highly favourable results. Allowing decision-makers to minimise costs while keeping the probability of poor checkpoint performance to acceptable levels.