Abstract: The Junggar Basin is an important contributor to reserves and production increase in China. To enhance deep-zone seismic imaging in the central basin, Sinopec Shengli Oilfield has successfully developed vibroseis techniques for high-density acquisition in various surface conditions, e.g. deserts, farmlands, and rivers. Seasonal frozen soil layers in near-surface desert areas caused by low temperature and precipitation in winter would affect the quality of raw vibroseis data. To investigate their influence and relevant countermeasures, we accomplish a detailed study of frozen soil layer structures, synthetic responses of geologic models, and field seismic responses based on the formation mechanism of seasonal frozen soil layers in desert areas. The study reveals that frozen soil layers affect vibrator-earth coupling effect, resulting in weakened downward propagation energy, moderately improved signal-to-noise ratio at medium to high frequencies. The countermeasures of employing sweep signals based on an optimized damped Ricker wavelet for excitation combined with receiving below the frozen soil layer significantly improve the imaging quality of deep-zone data. The research results provide a solution for improving vibroseis data quality under similar operating conditions.