Abstract: Soil salinization significantly impacts agricultural productivity and represents a major global issue in soil degradation. Timely and accurate monitoring of soil salinization dynamics is vital for the amelioration of saline-alkali lands. In recent years, remote sensing technology has played an increasingly crucial role in regional soil salinity monitoring. The use of salinity/vegetation remote sensing indices for soil salinity inversion has been prevalent. However, the accuracy of these indices is often compromised due to variations in geographic environments and crop cover conditions, leading to substantial uncertainty in inversion results. This study, set in the Yellow River Delta region in Shandong Province, China, aims to enhance the precision of regional soil salinity remote sensing inversion by integrating multi-temporal Sentinel-1/2 data with novel remote sensing indices adapted to different crop cover conditions. Employing multi-temporal Sentinel-1/2 remote sensing data coupled with concurrent ground sampling, this research focuses on analyzing the sensitivity of existing soil salinity inversion indices. It involves constructing new remote sensing indices specifically designed for different types of crop covers. The methodology encompasses a detailed correlation analysis between existing soil salinity data and 35 remote sensing spectral indices. This analysis is instrumental in identifying indices with a higher correlation to soil salinity, such as the Salinity Index 2 (S2), Normalized Differential Salt Index (NDSI), Canopy Response Salinity Index (CRSI), and Enhanced Normalized Differential Vegetation Index (ENDVI). Further, the study involves filtering through salinity big data remote sensing combinations of bands2、3and4, leading to the proposal of novel salinity indices–SI_YRDveg for crop-covered periods and SI_YRDbare for bare soil conditions. The research findings are notable in several aspects. (1) A correlation analysis conducted using existing soil salinity data and 35 remote sensing spectral indices has identified four remote sensing indices with higher correlations with soil salinity. These include the “Blue-Near Infrared” salinity index (Salinity index 2, S2), the Normalized Differential Salt Index (NDSI), the Canopy Response Salinity Index (CRSI), and the Enhanced Normalized Differential Vegetation Index (ENDVI). (2) By screening various combinations of 2, 3, and 4-band indices for soil salinity data, novel salinity indices were proposed for crop cover (SI_YRDveg) and bare soil (SI_YRDbare) conditions. The novel indices achieved the highest correlation of 0.77 across multiple temporal phases, significantly enhancing the correlation between remote sensing indices and soil salinity. (3) Using the novel salinity indices combined with the Random Forest method, the salinity inversion accuracy for crop-covered surfaces achieved an RMSE of 2.23 g/kg and an MAE of 1.25 g/kg, while for bare soil surfaces, the RMSE and MAE were 1.00 g/kg and 0.73 g/kg, respectively. Compared to inversions using only basic remote sensing bands, accuracy improved by 31%.This study underscores the effectiveness of integrating advanced remote sensing indices and machine learning techniques to enhance the accuracy of soil salinity mapping, thereby providing more reliable data for agricultural management and environmental monitoring. The study provides a novel set of salinity indices and an improved methodology for soil salinity inversion in regional cultivated lands, particularly in areas with diverse vegetation covers. The integration of Sentinel-1/2 data with these newly developed indices significantly enhances the precision of soil salinity monitoring. This advancement is a crucial step forward in remote sensing applications, offering a more reliable and nuanced approach to understanding and managing soil salinity in agricultural landscapes. Bare soil surface shows an increasing trend in soil salinity due to evaporation and salt migration, while vegetated surfaces suppress salt accumulation by reducing evaporation, enhancing moisture retention, and increasing salt absorption. The success of this method highlights the potential for further development and application of tailored remote sensing indices in addressing complex environmental challenges in agriculture.