A modified cellular automaton (CA) model based on the NaSch model is proposed to simulate the typical reckless jaywalking which can cause serious traffic jam and even traffic accident. Incorporating traffic psychology, the walking characteristics such as customary activity, randomness, conformity and tendency are analyzed. Two parameters reflecting the influence degree of human factor is introduced with the global consideration of the above walking characteristics. The numerical simulations are carried out under the periodic boundary conditions. Different cases including the reckless jaywalking at a fixed position and other stochastic positions are taken into account, respectively. The relationship between the traffic states and weight of human factor is investigated. The corresponding fundamental diagrams and the spatio- temporal patterns are given. Furthermore, the formation mechanism of traffic jam is explored and the measures for decreasing the negative influence of human factors are discussed.