Monitoring and understanding the dynamics of forest ecosystems is vital for the sustainable management of forest resources and assessing potential risks of climate change. However, the spatial-temporal variability within the ecosystem and rapidly changing forest inventory and management policies makes forest mapping a complex process. Advancements in remote sensing (RS) technologies over the years have assisted in improved mapping and retrieval of forest parameters over regional and global scales. Synthetic Aperture Radar (SAR) technologies have revolutionized the role of RS in forestry applications due to its all-weather, self-illumination, and penetration capabilities. With numerous current and future SAR sensors, a high volume of data will be available, serving as the workhorse for forest monitoring. In recent times, machine learning (ML) approaches have shown great ability to self-learn and predict for a different range of RS applications. Herein, the importance of ML for tackling different forestry applications employing SAR data will be discussed. The unique feature of SAR technology along with the most relevant ML models, will be reviewed. Subsequently, different classification and regression problems with illustrative examples are presented to demonstrate the ML capability in forestry. Finally, a conclusion regarding the open challenges and future direction is presented.