Smart Forestry 5.0 in conjunction with machine learning-driven optimisation is revolutionising forest management by enhancing resilience, sustainability, as well as efficiency. Predicting how climate variation will affect forests is crucial for proactive decision-making as the threat to global ecosystems grows. Advanced machine learning techniques improve carbon capture, forest development, as well as biodiversity preservation by making precise forecasts of environmental changes. However, it is challenging to predict how climate variation may affect intelligent forestry. Reliable projections are made more difficult by the complex nature related to climatic elements, the unpredictability associated with biological information, as well as the dynamic nature linked with forest ecosystems. Additionally, robust computational methodologies as well as adaptable methods are required for integrating remote sensing, Internet of Things (IoT), as well as big data analytics. Interdisciplinary approaches that combine Artificial Intelligence (AI), climate science, as well as sustainable forestry practices are needed to overcome these challenges. Hence, this chapter performs the Climate Change Prediction (CCP) model on Smart Forestry 5.0 with the help of innovative machine learning-based optimisation methodology. The data pertaining to the climate change on forests is first gathered from the sixth phase of the Coupled Model Intercomparison Project (CMIP6). The pre-processing of these gathered data is then accomplished using the Kalman filtering approach. From these pre-processed data, the extraction of the features is done by the Hilbert-Huang Transform (HHT) approach. Finally, the prediction of these extracted features is performed by the novel machine learning model called Enhanced Streaming Gradient Boosted Trees (ESGBT), where the parameter tuning of SGBT is accomplished by the optimisation algorithm referred as Wolverine Optimisation Algorithm (WoOA). Prediction accuracy maximisation and error minimisation is returned as the final fitness function. The findings demonstrate that the proposed ESGBT-WoOA for the CCP model on Smart Forestry 5.0 surpasses the other methods in terms of distinct error measures. The proposed ESGBT-WoOA for the CCP model on smart forestry is 13.06% and 61.65% better than the other state-of-the-art methods in terms of prediction accuracy and error, respectively.