This paper presents a model based on Queueing Theory for Orthogonal Frequency-Division Multiplexing-Time-Division Multiple Access (OFDM-TDMA) networks. This network model requires a Markovian model to the traffic processes. Thus, we present a Markov Modulated Fluid Model (MMFM) with autocorrelation function fitting for the network traffic, in order to evaluate the performance of the transmission link in OFDM-TDMA systems. That is, we deduce equations for estimating QoS parameters of OFDM-TDMA systems. The obtained results confirm that the proposed network model is efficient in describing the OFDM based link performance. A novel equation is also introduced for estimating the buffer overflow probability in OFDM-TDMA systems using a proposed network service curve. The computational effort is reduced compared to the Queueing Theory based method of overflow probability estimation. Besides, the results show that the estimates are very close to those obtained by simulations Ver menos