Few cellular automaton models of traffic have taken the relationship between the randomization probability of velocity and the specific value of current velocity into account. On the basis of Nagel-Schreckenberg model, this paper presents a modified cellular automaton model in which the randomization probability evolves with the driver’s remembering to historic experience and with the traffic condition around. Numerical simulation results show that the proposed model can reveal the influence of driver’s remembering on the traffic and generate higher average velocity and flux.