<p>Big data analytics is transforming smart agriculture by enabling data-driven decision making to optimize operations, enhance productivity, and minimize the usage of resources. Moreover, effective agricultural management relies on analyzing the moisture levels of soil, weather conditions, and water requirements for the crops while also providing early warnings for pest infestations and disease outbreaks. Additionally, big data analytics supports environmental monitoring by tracking soil quality, water pollution levels, and carbon emissions, promoting sustainable farming practices. To address the challenges of both quality and quantity of crop production, this paper proposes an Attention Bidirectional Convolutional-based Random Light Spectrum Algorithm. This proposed model integrates a Convolutional Neural Network for spatial agricultural data analysis, allowing for accurate detection of crop diseases, growth patterns, and pest infestations. In addition, the Attention-based Bidirectional Long Short-Term Memory networks process histories and sequential data, for instance, previous irrigation practices or environmental history, while the attention mechanism enhances the focus on critical data points. Furthermore, hyperparameters are optimized by metaheuristic optimization techniques. The validation of experimental results on the Crop Production in India dataset demonstrates the superior performance of the Attention Bidirectional Convolutional-based Random Light Spectrum Algorithm. Comparative analysis against existing techniques highlights its effectiveness, achieving 98.5% accuracy, 97.6% F1-score, 97.2% precision, 96.9% sensitivity, and 96.3% specificity. These results affirm the algorithm's potential to enhance crop yield and quality, providing a robust, efficient, and sustainable solution for modern agricultural challenges.</p>