Next-generation communication systems are expected to be highly integrated heterogeneous networking environments. Supporting multimedia services with different quality-of-service (QoS) requirements over such highly heterogeneous environments is a very challenging task, and a unified call admission control (CAC) optimization mechanism is needed to allow resources to be efficiently utilized. In this paper, we propose an optimal CAC mechanism, i.e., the guard-channel-basedincrementalanddynamicoptimization (GUIDO), which is a threshold-based CAC scheme that is designed based on the Markov decision process (MDP) and sensitivity analysis (SA) techniques. The MDP is used to derive the optimal CAC policy for predetermined goals, whereas the use of SA can greatly reduce the effort of unnecessary recomputation of the CAC policy when the traffic condition varies. Through the use of system capacity estimation and QoS mapping techniques, GUIDO allows the CAC operation to be performed in a general framework, even with various underlying communication systems and different access/core networks. GUIDO can be implemented dynamically (online optimization) and statically (lookup table). Homogeneous and heterogeneous systems with multiple service priority classes are studied in this paper. We have considered comprehensive objective functions, and each characterizes a different interest of the service provider. The performance is evaluated via an event-driven simulator, and the results are compared with those of the complete sharing scheme and conventional guard channel schemes under constant call-arrival rates as well as time-varying call-arrival-rate traffic scenarios.