Open Access

REVIEW

Systematic Review of Machine Learning Applications in Sustainable Agriculture: Insights on Soil Health and Crop Improvement

Vicky Anand¹, Priyadarshani Rajput¹, Tatiana Minkina¹, Saglara Mandzhieva¹, Santosh Kumar², Avnish Chauhan³, Vishnu D. Rajput^1,*

1 Academy of Biology and Biotechnology, Southern Federal University, Rostov-On-Don, 344090, Russia
2 Department of Geography, Shaheed Bhagat Singh College, University of Delhi, New Delhi, 110017, India
3 Department of Environment Science, Graphic Era Hill University, Bharu Wala Grant, 248002, India

* Corresponding Authors: Vishnu D. Rajput. Email: ,

(This article belongs to the Special Issue: Integrated Nutrient Management in Cereal Crops)

Phyton-International Journal of Experimental Botany 2025, 94(5), 1339-1365. https://doi.org/10.32604/phyton.2025.063927

Received 29 January 2025; Accepted 16 April 2025; Issue published 29 May 2025

Abstract

The digital revolution in agriculture has introduced data-driven decision-making, where artificial intelligence, especially machine learning (ML), helps analyze large and varied data sources to improve soil quality and crop growth indices. Thus, a thorough evaluation of scientific publications from 2007 to 2024 was conducted via the Scopus and Web of Science databases with the PRISMA guidelines to determine the realistic role of ML in soil health and crop improvement under the SDGs. In addition, the present review focused to identify and analyze the trends, challenges, and opportunities associated with the successful implementation of ML in agriculture. The assessment of various databases clearly revealed that ML implementation depends on crop management, while its limited potential in terms of soil health was explored. ML models, such as random forest and XGBoost, have demonstrated high accuracies of up to 99% in crop yield prediction and disease detection. Advanced ML frameworks, including the SHIDS-ADLT and EfficientNetB3, have improved soil health monitoring and plant disease classification. Irrigation management using ML has achieved over 50% water savings and irrigation efficiency by 10%–35%. These findings highlight the potential of ML to improve sustainable agricultural practices and soil health. A significant improvement discussed in this review is AutoML, which simplifies ML model implementation by automating feature selection, model selection, and hyperparameter tuning, reducing dependency on ML expertise. The integration of ML with remote sensing, Internet of Things (IoT), and big data analytics is expected to further transform the precision agriculture and real-time decision-making approaches to optimize resource utilization. Conclusively, the present review offers a quantitative perspective on the evolution of ML in agriculture, soil health management, crop yield prediction, and resource optimization.

---

Keywords

---