<p>The impact of the agricultural drought on rain-fed crop growth significantly impacts employment possibilities and per capita income. Agriculture drought is a significant problem across nations. It is predominantly a climate condition that cannot be avoided. Satellite imagery produces timely and consistent spatial information. Decision-makers can successfully manage agricultural resources by monitoring crop quality with the help of spatio-temporal information. Triple EXponential - Long Short-Term Memory Model (TEX-LSTM) is proposed for agricultural drought prediction with less Root Mean Square Error (RMSE). In the present analysis, MODIS sensors 13Q1 and 11A1 are used to acquire the data in a timely mode. The collected data set is processed for Time Series Analysis (TSA) to ascertain the underlying reasons for long-term trends or systemic patterns. The time series data is inputted to the triple exponential smoothing technique to extract the tuned parameters, such as magnitude, repeating, and shifting factors. When smoothing techniques are used, customized parameters are extracted and applied to standardized time series data using the LSTM algorithm. The drought intensity throughout the times and regions is determined using the vegetation index to examine a specific duration. The proposed TEX-TSTM model is trained and evaluated using Standard Vegetation Index (SVI) and Vegetation Health Index (VHI) datasets with a 250-meter resolution from 2000 to 2020. The performance of the proposed system is analysed in terms of R2_Score, RMSE, Mean Absolute Error (MAE), and Mean Square Error (MSE), and TEX-LSTM is found to yield the highest validation accuracy of 95.27%. Hence, it can be used as a computer-assisted technique to guarantee sufficient water resources at the right moment to promote crop growth, increase agricultural yields, and meet food resource requirements.</p>