Cognitive Radio(CR) system based on Orthogonal Frequency Division Multiple Access(OFDMA),such as Wireless Regional Area Networks(WRAN) and Worldwide Interoperability for Microwave Access(WiMAX),often attempt to improve performance via dynamic radio resource management,which is characterized as concurrent processing of different traffic and nondeterministic system capacity.It is essential to design and evaluate such complex system using proper modeling and analysis tools.In the previous work,most of the communication systems were modeled as Markov Chain(MC) and Stochastic Petri Nets(SPN),which have the explicit limitation in evaluating adaptive OFDMA CR system with wide area traffic.In this paper,we develop an executable top-down hier-archical Colored Petri Net(CPN) model for adaptive OFDMA CR system,and analyze its performance using CPN tools.The results demonstrate that the CPN can model different radio resource manage-ment algorithms in CR Systems,and the CPN tools require less computational effort than Markov model using Matlab,with its flexibility and adaptability to the traffics which arrival interval and processing time are not exponentially distributed. Cognitive Radio (CR) system based on Orthogonal Frequency Division Multiple Access (OFDMA), such as Wireless Regional Area Networks (WRAN) and Worldwide Interoperability for Microwave Access (WiMAX), often attempt to improve performance via dynamic radio resource management, which is characterized as concurrent processing of different traffic and nondeterministic system capacity. It is essential to design and evaluate such complex systems using proper modeling and analysis tools. In the previous work, most of the communication systems were modeled as Markov Chain (MC) and Stochastic Petri Nets (SPN), which has the explicit limitation in evaluating adaptive OFDMA CR system with wide area traffic. This paper, we develop an executable top-down hier-architected Colored Petri Net (CPN) model for adaptive OFDMA CR system, and analyze its performance using CPN tools. The results demonstrate that the CPN can model different radio resource manage-ment algorithms in CR Systems, and the CPN tools require less compu tational effort than Markov model using Matlab, with its flexibility and adaptability to the traffics which arrival interval and processing time are not exponentially distributed.