The step over plays a significant controlling role in the propagation process of seismic rupture. Researchers, based on geological survey results of multiple strike-slip earthquakes globally, have indicated that it is challenging for seismic rupture to cross step over wider than 5 km. Results from numerous numerical simulations also demonstrate that seismic rupture, under conditions such as uniform elastic media and uniform stress fields, cannot extend beyond step over wider than 5 km. However, as research progresses, it has been observed that in few earthquakes, rupture does cross step over wider than 5 km. And research fellows offer various explanations for this phenomenon, including higher stress near step over, changes in material properties due to rock damage, or the presence of concealed faults, among others. Drawing inspiration from previous studies, this paper proposes a novel explanation. In numerical simulations of the seismic rupture process, it is commonly assumed that rupture occurs only on pre-existing faults. Nevertheless, it is highly likely that new ruptures may generate outside pre-existing faults during seismic events, giving rise to new faults. The rocks near step over are relatively brittle and are more likely to generate new ruptures during seismic events, thereby allowing rupture to cross wider step over. This study employs the extended finite element method to investigate the influence of newly generated ruptures near step over on the seismic rupture propagation process. A step over model with a width of 10 km is established. The newly generated ruptures in the model follow the maximum shear stress failure criterion; when the maximum shear stress exceeds the rock’ s ultimate limit, a new fault is formed in the intact medium, and the direction of the new rupture’ s expansion corresponds to the direction of maximum shear stress. According to the results of rock physical experiments, the model medium’ s maximum shear stress limit is set as 72 MPa. The simulation results using the extended finite element method in this study show that after an earthquake occurs, the rupture first propagates to the end of the pre-existing fault, and then new ruptures occur near the step over. The expansion direction of the new rupture is nearly perpendicular to the strike of the pre-existing fault, and the fault continues to expand for 1.8 s, with an expansion length of approximately 8.2 km and an expansion velocity of approximately 4 556 m/s. This study also establishes a model that does not consider newly generated ruptures for comparison. Without considering new ruptures, the rupture cannot cross a 10 km wide step over. The paper compares Coulomb stress distribution maps at a specific moment in these two scenarios. According to the Coulomb stress distribution, it is evident that without considering newly generated ruptures, the rupture in this model can only cross step over for the model approximately 5 km wide. The new ruptures alter the geometry of the step over and greatly change the spatial distribution pattern of Coulomb stress, especially by elevating stress levels on the fault, thereby enhancing the ability of seismic rupture to cross step over and allowing it to cross step over 10 km wide. This study contributes to a deeper understanding of the propagation process of seismic rupture across step over, particularly for seismic source process analysis and seismic hazard assessment.