<p>The analysis of power spectral densities of turbulence relative to moving vehicles is essential for assessing the crosswind stability of vehicles. To increase the accuracy and efficiency of these assessments, particularly in avoiding the disturbances introduced by complex flow structures around railways, a novel approach for measuring wind characteristics and calculating spectra relative to moving vehicles is proposed. This approach integrates a stationarity test, a quasi one-parameter equation, and a forward analysis method. Field measurements were conducted in a complex terrain adjacent to a railway using this proposed method. Wind characteristics were analysed based on the nonstationary process (through the stationarity test) to identify the nonstationarity (detected in 99.9% of the longitudinal wind speed fluctuations) and the correlation of wind with the complex terrain. The stationarity test is found to be more effective than a randomness test for isolating time-varying mean wind speed in such terrains, enhancing the accuracy of fluctuating wind characteristics detection by nearly 40%. The calculated ratios of lateral to longitudinal turbulence intensities and turbulence length scales are 0.972 and 0.418, respectively, with the lateral spectrum parameter being 0.579. These findings challenge the full applicability of isotropic turbulence models. The quasi one-parameter equation significantly reduces the effort required to capture wind fluctuations, while the forward analysis method, accounting for turbulence correlation, enables precise and efficient calculation of spectra relative to moving vehicles, improving accuracy by 12.4%. The methodology not only facilitates the accurate and effective acquisition of wind characteristics and spectra relative to moving vehicles but also lays the groundwork for further research on the crosswind stability of vehicles in real turbulent wind and for measuring wind characteristics in complex terrains.</p>