The neuromorphic photonic information processing technology is inspired by the high-speed and low power-consumption information processing mechanism in brain. Compared with the traditional photonic information processing technologies, it not only promotes greatly the processing speed and energy efficiency, but also possesses important applications in detection and perception. This paper proposes a neuromorphic photonic information processing scheme based on the distributed feedback(DFB) laser, which is used for high-speed motion direction selection and recognition. The influence of bias current on the DFB laser pulse response duration is analyzed. The dependence of high-speed motion direction recognition on the DFB laser pulse duration is confirmed and the module for high-speed motion direction selection based on DFB laser is constructed. The experimental test of this module is conducted. On the premise of correct recognition, the effects of link weighting parameters and current bias on the applicable speed range are analyzed. The results show that the proposed module can realize the effective recognition in one-dimensional direction within the Mach-level speed range.