Adverse weather is a leading cause in the disruption of transmission and distribution of electricity. A smart microgrid (SMG) seeks to improve a grid’s readiness in inclement weather through localization of renewable energy sources, use of smart sensors/meters, and use of an improved energy management system. This paper seeks to justify the implementation of SMGs and the importance of load forecasting in adverse weather with methodologies for hardware and software. Historical weather and electric load data were gathered, compared, and used to train a neural network. A scaled-down SMG was designed to capture power generation from renewable energy sources and forecast future generation outputs. An important aspect within these systems is the Internet of Things (IoT) which allows for wireless capturing and monitoring of data in real-time. Arduino Cloud was used to obtain real-time analytics from renewable energy sources. Finally, this data was modeled to fit in the previously built neural network to forecast future load amounts for the system. The correlation found between adverse weather and load generation, coupled with the ability to accurately forecast such load generation, highlights a need for improved grid stability and the acquisition of accurate, timely predictive grid metrics.