The degradation of ecological health in arid and semi-arid regions such as Saudi Arabia poses significant environmental and socio-economic challenge. This study aims to assess and model the ecological health of Saudi Arabia using advanced geospatial techniques and statistical methods. The main objectives include extracting relevant environmental data using Google Earth Engine (GEE), applying fuzzy logic to model ecological health, and using Ordinary Least Squares (OLS) regression to identify influential factors. The methods used include the integration of multiple data sets from MODIS and VIIRS, which include parameters such as nighttime light (NTL), aerosol concentration, carbon monoxide (CO), nitrogen dioxide (NO2), ozone (O3), sulfur dioxide (SO2), humidity, land surface temperature (LST), NDVI, NDWI, and precipitation. These parameters were processed and analyzed using GEE to create a fuzzy logic-based model of ecological health. The model identified five different ecological health zones: Very Good (191,444.14 km(2)), Good (77,856.82 km(2)), Moderate (50,521.63 km(2)), Poor (21,601.98 km(2)) and Very Poor (6,592.54 km(2)). The quantitative results of the OLS regression showed that NTL, CO and SO2 have a negative impact on ecological health, while humidity, NDVI, NDWI and precipitation contribute positively. NO2 in particular showed a strong positive influence, which could indicate managed urban environments with better ecological practices. The OLS model performed exceptionally well with an R-squared value above 0.9, indicating high explanatory power and reliability. Therefore, this study highlights the benefits of integrating big data analytics with fuzzy logic and statistical modeling to assess ecological health. The results provide a comprehensive understanding of the factors influencing environmental health in Saudi Arabia and offer valuable insights for targeted environmental management strategies. Future research should incorporate higher resolution datasets and ground-based observations to further refine the model and validate the results.