Abstract: A three-node Amplify-and-Forward (AF) cooperative diversity network model is proposed in the paper. The randomness of feedback delay analysis is presented for the proposed model which the source transmits a frame over Nakagami-m fading channel. The frame can be transmitted from the direct link or relay link to the destination node. Additionally, Adaptive Modulation and Coding (AMC) technique is used in the source node of the analytical model which allows multi-rate transmission of the data frame, the transmitted mode is chosen according to the channel conditions. Based on the above reasons, the steady distribution of the feedback delay is analyzed, and a Finite State Markov Channel (FSMC) is developed to evaluate performance the Nakagami-m channel and obtain the channel state transition matrix. Then the transition probability of the feedback delay is derived. The simulative results are compared with the theoretic results, the results verify that the theoretic analysis is correct. Finally, the effect of channel fading coefficients and maximum frame error rate on the probability distribution of the feedback delay performance is analyzed.