Most Stars are thought to be formed inside filamentary molecular clouds. Recent observations have proposed that star formation may be triggered through collisions between filamentary molecular clouds (Duarte-Cabral et al. 2010; Nakamura et al. 2014). Furthermore, according to Kumar et al. (2020), all luminous clumps are located at the intersections (referred to as "hubs") of filaments and this indicates the formation of massive stars preferentially form at the hubs. Based on the above, filament collisions are considered to be a universal and important phenomenon that is relevant from near- to massive star-forming regions. Understanding the filament collision process will reveal the initial conditions for collision-induced star formation. Therefore, we have been working on 2D-MHD simulations of filament-filament collisions. In this conference, I will report on the condition of radial instability of the merged filament and its evolution. As the initial condition, we prepare two identical infinite long filaments, threaded by lateral magnetic fields. The filaments are in magnetohydrostatic equilibrium(Tomisaka 2014; Kashiwagi & Tomisaka 2021). The two filaments collide head-on along the magnetic field lines with relative velocities ranging from transonic to supersonic. The main results are two: (1)Radial collapse model: The merged filament collapses when the initial total line mass exceeds the critical line mass and confirms that neither the strength of the magnetic field nor the initial velocity has an effect on the instability of the shocked region. (2)Stable model: When the initial total line mass is less massive than the critical one, a stable filament is formed with oscillating, and its density profile is very similar to that of the magnetohydrostatic equilibrium one. In addition to this, the results for the case where the collision direction is perpendicular to the magnetic field lines are also briefly introduced.