Soil moisture is a basic factor influencing agricultural productivity, water resources management and climate control. Proper estimation of soil moisture is essential to counteract the negative impacts of extreme weather events like drought and floods. Although in-situ measurement methods are highly accurate, they lack scalability. The remote sensing, especially the Sentinel-1 Synthetic Aperture Radar (SAR) data, has proved to be a strong candidate due to soil moisture, which changes despite climatic conditions or times. The research uses Google Earth Engine (GE) to process the Sentinal-1 SAR time-series data. The SAR images have been analysed using three deep learning methods such as Artificial Neural Network (ANN), Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) for estimating soil moisture. Preprocessing techniques are used to improve data quality such as radiometric calibration, speckle filtering and normalization of the area. Soil Moisture Index (SMI) has calculated from backscatter values (VH band), model training and performance measurement by accuracy, precision recall and F1-score. The results of our study showed that the RNN model performed better (96.12% accuracy) than the other models used. This model successfully shown the temporal dynamics in soil moisture and potential for its utility for large-scale monitoring purposes. These results showcase the advantages of combining deep learning methods to evaluate scalable and efficient soil moisture estimation which in turn will be utilized for water resources management and agricultural planning.