Abstract: The void distribution of carbon fibre reinforced plastics (CFRP) has an important effect on its mechanical properties. Focusing on this problem, recurrence quantification analysis (RQA) of ultrasonic back-scatter signals was proposed to characterize void size and position distribution. Three groups of CFRP models with spherical voids were established, in which each model had different void position distribution. The void diameters D ranged from 26 μm to 70 μm, and the distance d of voids was 0.21 mm, 0.14 mm and 0.09 mm, respectively. Based on these models, the back-scatter signals were analyzed by RQA method. The simulated study shows that, recurrence rate P_RR decreases with D increasing when the void position distribution is the same. If the void size distribution is the same, P_RR increases with d decreasing. The P_RR of models with d=0.09 mm are always higher than the other two groups. There is significant difference in P_RR between d=0.14 mm and d=0.21 mm models when D<50 μm. But the difference is small when D ≥ 50 μm. The results demonstrate that there is difference in P_RR on different void distributions, P_RR can be used to characterize different void distributions in CFRP.