Today's agriculture industry makes extensive use of the promising field of machine learning (ML). There is not enough labour available for agriculture, and there are not enough skilled farmers. It is difficult to identify and stop crop diseases without a thorough understanding of the current situation. It is also frequently used in many aspects of agriculture, including managing soils, yields, water, diseases, and weather. The ML models allow rapid and actual decision-making. To anticipate correctness of the output, ML model uses training and testing. Species management, Disease detection, yield prediction, crop quality, water management, weed identification, increased productivity and better management of soil categorization are all aided by an application of ML in agriculture. By highlighting benefits and drawbacks of various ML methodologies put forth in the last five years, this article seeks to provide comprehensive information on them. Additionally, it contrasts several ML techniques employed in contemporary agriculture.