Abstract: The exploration areas in western piedmont zones have emerged as an important strategic replacement area for hydrocarbon exploration. However, the complex topography and structure ("dual-complexity"), along with the seismic data with low signal-to-noise ratios (SNR), severely restricted the accuracy of seismic imaging and the success rate of drilling. Therefore, it is imperative to optimize inversion technologies. This study focused on the western piedmont zone, an area characterized by complex surface geology, including hills, gravelly zones, and yardang landforms, further degrading the quality of seismic data. The existing first-arrival waveform inversion technique relies on first-arrival traveltime fitting. However, due to the near-surface heterogeneity in the western piedmont zones, this technique faces two core limitations: insufficient wavelet consistency and insufficient surface consistency. This results in inadequate accuracy of shallow velocity modeling and makes it difficult to invert the velocities of intermediate-to-deep layers using ultra-large-offset seismic data and the abundant low-frequency information contained in first-arrival waves. In contrast, reflection waveform inversion (RWI) exhibits higher adaptability to low-quality data by exploiting information of intermediate-to-deep layers in reflected waves and optimizing the objective functions accordingly. Based on this, this paper applied RWI to the survey area, aiming to enhance the accuracy of migration velocity inversion under dual-complexity conditions and to provide technical reference for improving seismic imaging quality and the detailed depiction of hydrocarbon reservoirs in western piedmont zones. Model and field tests demonstrate that the RWI can handle low-quality data in piedmont zones and improve the overall accuracy of velocities in intermediate-to-deep layers.