A novel approach to using tunable diode laser absorption spectrum (TDLAS) is developed for nonuniform velocity distribution measurement by Doppler effect. An analysis of the energy in direct absorption spectrum at low frequencies is made by Fourier transform, because the TDLAS method offers the advantages in using Beer law to deal with coupling relations between velocity distribution and corresponding length of velocity region. By comparing with traditional TDLAS-Doppler velocity measurement, advantages of this approach to the more exact solution of core flow velocity by signal process without using extra lasers and detectors are explored. Following the published theory, between velocity regions at multiple projections the absorbance about average in frequency offsets and the absorbance about difference in frequency offsets are incorporated into an improved fitting model. A solution to obtaining changes of absorbance energy at low frequencies by Fourier transform is used to demonstrate the ability to recover minor change in absorbance under different conditions, inferring a better method to realize the simultaneous measurement of velocity distribution. The influences of these parameters, such as projection angles and noise during absorption, are investigated by the multiple projection simulations at rovibrational transitions of H₂O near 7185.6 cm^–1 from three projections. This approach is validated in a two-stage velocity distribution model, demonstrating the ability to exactly measure core flow, with a precision of 0.9% RMS (root mean square). The high velocity in the core flow is less influenced by the random noise in absorption due to nearly linear relationship between the difference in frequency offsets and the ratio of length of velocity region. Some satisfied results can be obtained when larger angles of projection are arranged. The combination of 0°, 30°, and 60° will be a reasonable optic design considering the limitation of spatial resolution. In conclusion, the novel approach to velocity distribution measurement based on TDLAS-Doppler from multiple projections has great potential applications in engine diagnosis and gas dynamic research.