Accurate characterization of soil water-salt distribution is critical for sustainable saline soil management in arid regions. This study systematically evaluates three spatial interpolation methods, i.e., Ordinary Kriging (OK), Co-Kriging (CK), and Inverse Distance Weighting (IDW), in the Yarkand River Basin of Northwest China using 100 topsoil (0-20 cm) samples. OK was used here as the baseline geostatistical method, providing best linear unbiased estimates and uncertainty quantification while tending to over-smooth local variability. A combined approach integrating IDW-derived covariates into CK was developed to mitigate OK's over-smoothing, while additional random forest optimization further enhanced the prediction robustness. Hold-out validation with 80 training points and 20 testing points revealed that OK and CK have comparable accuracy (with correlation coefficients (R) of 0.48 and 0.49 for soil salt and 0.17 and 0.16 for soil water content, respectively), and they are both superior to IDW (with R of 0.19 for soil salt and 0.10 for soil water content). CK outperformed OK in cross-validation with R improved from about 0.35 to over 0.80, demonstrating its sensitivity to localized variability. The combined method increased R of test datasets by 319 % and 49 % for soil water content and total salt content, respectively. This combined method is applicable for soil properties with different spatial heterogeneity, as indicated by the coefficients of variation of 38.7 % for soil moisture content and 93.9 % for total salt content. Interpolation results indicate that the soil water content is 0.04-0.28 g/g in the oasis with no obvious trend in the overall distribution. The total soil salt content is 0.87-12.1 g/kg with obvious spatial heterogeneity, i.e., nonsalinized in the upstream, slightly salinized in the midstream, and moderately salinized in the downstream. These findings establish a transferable framework for multi-method integration in spatial interpolation of soil properties, which balances algorithmic strengths with environmental heterogeneity.