Energy is critical to the social and economic development of any country. The increase in industrial and economic activities, energy demands have grown rapidly. Hence energy challenges pose to become a great concern. Ever-increasing growth in energy demand throughout the world, traditional resources’ (fossil fuel) reduction, and high rate of pollution and the resulting imposed costs have led the energy producers to move toward harvesting renewable energy (Soman et al., 2010). Over the decades, Windmills were used for power extractions from wind energy. The historic windmills made from wood, cloth and stone of typically large, heavy, and inefficient machines has now evolved to wind power devices. Wind power devices are presently used to produce electricity and are termed wind turbines. Wind energy, as the most environmental-friendly energy source, has attracted substantial global interest. Wind is infinite; some of the world best wind sites for wind farms installation are the cold- regions, because of the low-temperature conditions. In cold- regions, the air density is favourable for wind energy extractions because of these low-temperature conditions. Super-cooled droplets and precipitation affect the wind turbine operations and change the aerodynamic profile of the wind turbine blade with ice accretion. Ice loads on wind turbine blades led to uneven balancing on the leading edge deteriorates the aerodynamic characteristics of the turbine blades, and often the wind turbine is shut down. The ice accretes because of super-cooled water, causing air bubbles to be trap within the ice. Hence ensuing in a solid phase characteristic ice, still with an increase in temperature this ice further melts, and the droplets then impacts on the surface and flow over it before freezing to form ice (glaze). This paper considers the review of Machine learning (ML) models to the predict air density around wind turbine in cold and low-temperature conditions. Artificial neural networks (ANNs) are combined with a statistical weighted pre-processing method (SWP–ANN) were used to predict airflow behavior, droplet behavior, and resultant ice accretion on wind turbine. The study shows that Machine learning can monitor the ice accretion on wind turbine blades in cold regions. The wind turbine rotation speed is dependent on the temperature and air density around the turbine blade. This paper reviews the operation of wind turbines in humid and cold regions.