A wide variety of algorithms are used in machine learning to develop models that can anticipate future data that has not yet been observed and learn predictive rules from existing data. Using machine learning techniques in agriculture is one of the emerging topics in recent research fields. The main objective of smart agriculture is to increase profit by lowering agricultural costs. Smart farming is made possible by machine learning algorithm's predictive nature. To increase productivity, clever farmers use this technique in their agriculture. Soil is one of several important factors that need to be considered. The composition of soil nutrients, its texture and pH have serious impacts on the growth of crops, while many farmers are not aware of this. One of the most important criteria in agriculture is selecting the right crop for the soil conditions of the farmland. The main focus of this survey is on crop suggestions and crop recommendations. This paper provides a brief analysis of the use of machine learning techniques in crop recommendation systems based on soil and crop yield prediction systems.