Review of machine learning and blockchain integration for crop monitoring and management in smart farming

Abstract

Agriculture is an essential pillar of global food security, economic development and employment. However, this sector is facing major challenges such as population growth, soil degradation, water stress, climate hazards, and dependence on chemical inputs, which directly affect productivity, crop quality and the sustainability of natural resources. In this context, the integration of digital technologies and smart solutions appears as an essential step to modernize agricultural systems and strengthen their resilience. This review paper analyzes the role of Machine Learning (ML) and Blockchain as complementary technologies contributing to the transformation of traditional agriculture into smart, connected and secure systems. The study is based on a qualitative analysis of more than 127 recent scientific publications selected according to the PRISMA methodology, and from recognized databases such as ScienceDirect, Springer, IEEE Xplore, MDPI and other academic reference sources. The results show that ML is widely used for key applications such as yield prediction, early detection of crop diseases, irrigation optimization and agricultural input management. These applications are based on different learning paradigms, including supervised, unsupervised and reinforcement learning, capable of exploiting heterogeneous and dynamic agricultural data. Moreover, the blockchain plays a fundamental role in securing, tracing, and making agricultural data transparent. It ensures the integrity of information, while facilitating integration with IoT systems and smart contracts to strengthen trust between producers, distributors and consumers, and to promote more sustainable and equitable agricultural supply chains. This research also highlights the combined benefits of these technologies, including improved resource efficiency, increased productivity and support for sustainable agriculture. However, several limitations persist, notably the quality and availability of data, scalability issues, energy costs of blockchain infrastructures and the complexity of their integration into real agricultural environments. Finally, this study identifies the main research gaps and highlights the need to develop hybrid, robust and integrated solutions combining ML and Blockchain. This critical synthesis provides a conceptual and practical framework to guide future research towards smart, resilient and sustainable agricultural systems capable of addressing global food security and natural resource conservation issues.