Machine learning (ML) has become a pivotal element in agriculture, providing groundbreaking solutions to tackle intricate issues related to productivity, sustainability, and resource management. A comprehensive examination of the current literature is crucial as the discipline evolves, allowing for the identification of significant themes, trends, and focal discussions. The current study employs latent Dirichlet allocation (LDA)-based topic modeling to examine 1114 publications regarding ML applications in agriculture, sourced from the Scopus database. The analysis indicates notable expansion in ML studies, featuring leading publications across various interdisciplinary fields. Six primary areas have been identified: precision agriculture and remote monitoring, molecular and food composition analysis, food systems and agricultural applications, quality assurance and adulteration detection, advanced financial and technological applications in ML, and predictive modeling for agricultural success and efficiency. Every topic is examined to highlight its contributions and possible avenues for further investigation. The analysis offers theoretical perspectives on the interdisciplinary aspects of ML in agriculture, along with practical applications for farmers, agribusiness experts, policymakers, and technologists. This study represents the first thorough review of ML applications in agriculture utilizing the LDA approach. It provides a current and comprehensive understanding of the field, while also uncovering emerging areas and opportunities for future exploration.