Crop Yield Prediction Using Precision Agriculture and Smart Farming Technologies: A Systematic Review and Future Research Trends

The use of crop yield predictions is essential in today's agricultural systems as the global food system is facing increased pressures from climate change, soil degradation, and water scarcity. Precision agriculture is based on the development of data-driven solutions for yield forecasting using remote sensors, IoT (Internet of Things) devices, machine learning, and crop growth models. This systematic literature review synthesizes recent research (2020–2025) concerning crop yield prediction by reviewing what types of data are available, factors affecting yield, what predictive models have been developed, and how those models are used in field, regional, and large-scale agricultural situations. This review also demonstrates how climatic and hydroclimatic variability interact with soil properties and management practices to create yield outcomes, while evaluating the accuracy and performance of statistical, machine learning, deep learning, and hybrid modelling approaches. The study concludes that multi-modal data fusion, real-time data assimilation, and hybrid physics/AI modelling approaches improve prediction accuracy and robustness. Ongoing challenges include (but are not limited to) model transferability, the lack of standardized benchmark datasets for model development and evaluation, insufficient real-time integration of multiple types of information to assist in yield prediction, and difficulty in explaining how advanced AI models develop yield predictions. The authors propose a future research agenda based on the findings of this study that will provide a standard multimodal dataset for the development of explainable AI, a scalable edge/cloud architecture for model deployment, and models that can be transferred across crops, climates, and production systems.