AgroMind+: An Intelligent Multi-Crop Recommendation System With LSTM-Attention Architecture and Predictive Sustainability Index

Precision agriculture requires intelligent decision-support systems that account for temporal agricultural dynamics and sustainability constraints. Traditional crop recommendation models are based on static learning methods and they fail to consider seasonal patterns or environmental effects. In this research, AgroMind+, a multi-crop recommendation system based on an attention-augmented LSTM that utilizes temporal soil, climatic and historical agricultural data to produce ranked crop recommendations. A new Predictive Sustainability Index (PSI) is combined into the recommendation algorithm to assess water efficiency, nutrient use, yield potential, and climate stress. Further, an Adaptive Crop Advisory and Intelligence (ACAI) module is available to provide actionable agronomic advices. Experimental results indicate that AgroMind+ attains 94.81% accuracy for classification and 97.21% Top-4 recommendation performance, outperforming traditional machine learning and standard deep learning models. The attention mechanism also enhances interpretability by highlighting important covariates such as soil nutrients and rainfall patterns, where it reveals AgroMind+ standard’s effectiveness in precision agriculture with focus on sustainability.