In this paper four conflict prediction approaches are considered: a classical geometric approach, two variations of a probabilistic approach developed by Paielli ÖC Erzberger, based on conflict probability and overlap probability, and a novel probabilistic approach. The objective of all conflict prediction approaches is to evaluate a set of planned or predicted trajectories on their conflict potential and to supply other Air Traffic Management (ATM) subsystems with the conflict information. The classical geometric approach and approaches based on conflict probability and overlap probability are briefly reviewed. The novel probabilistic approach is described and explained in more detail. Simulation results for ATM examples are provided and compared for the four approaches on flexibility of usage and imposed restrictions on aircraft behaviour.