Abstract: Molecular Tagging Velocimetry (MTV) and Particle Imaging Velocimetry (PIV) are often used for flow visualization and velocity field imaging. However, the requirement for tracer particles can bring systematic errors to the velocity measurement of PIV method when the tracer particles have poor followability and uneven distribution. In the case of MTV, although particle seeding is not required, the finite fluorescence lifetime of the tracer molecules typically constrains its use only in high-speed and supersonic flows. To develop a velocity field imaging method with no requirements of tracer particles and suitable for low-speed flows, a novel MTV method based on infrared (IR) laser-induced fluorescence is developed and verified in axisymmetric turbulent jet of carbon dioxide. Resonant vibrational transition of the small gas molecule is selectively excited by an infrared pulsed laser to achieve molecular tagging, and the fluorescence distributions of the excited molecules at different instants are then imaged by an infrared camera, from which the velocity distributions are deduced. The effects of the molecular vibrational energy transfer process model, finite fluorescence lifetime, lateral velocity component and molecular diffusion motion on the fluorescence distribution are analyzed to improve the accuracy of the velocity measurement. The proposed method has been successfully verified in carbon dioxide turbulent jets with velocities ranging from 5 m/s to 51 m/s, and the radial distribution of the axial velocity in the main region of the jet is measured. The radial spatial resolution can reach 107 microns, and the velocity distribution is consistent with the theoretical calculation of turbulent jets and previous experimental results. The relative uncertainty of velocity measurement is better than 8%. This method can be used to obtain high-resolution instantaneous velocity imaging of low-speed flow field. Subsequently, by improving the pulse power, excitation efficiency and repetition frequency of the infrared laser, the measurement accuracy, spatial resolution and temporal resolution of this method can be further improved. Therefore, the proposed method bears great potential to provide a quantitative velocity field imaging method in the near-wall flow, micro-scale flow and large gradient flow where it is difficult to introduce tracer particles.