Establishing predictive models for the embodied and operational carbons is vital to facilitate the quantification of the life cycle emissions of a green building. However, embodied carbon is often overlooked compared to the operational predictive model due to the complexity of collecting embodied data. Previous studies relied on basic parameters (i.e., GFA, storey, etc.) to quantify embodied carbon. This approach leads to uncertainty due to variations in building materials. Identifying material-specific parameters that contribute to the most embodied emissions is crucial to enhancing the prediction accuracy of regression models. Taking Malaysia’s green-certified office buildings as a case study, this study aims to develop an integrated Life Cycle Assessment-Multiple Linear Regression (LCA-MLR) framework to establish hotspot-oriented regression models. Results show that cement and steel contribute two-thirds of embodied emissions and are adopted in the MLR model to predict embodied carbon. Hotspot-based embodied regression equations, formulated using material-specific parameters, enable accurate prediction with 23.7–42.1% variation compared to actual embodied emissions, while gross floor area and storey incur 90.6–93.4% and 99.7–99.9% variation, respectively. Using the hotspot-oriented regression models, a new Green Office Building Carbon Rating is established to benchmark the carbon performance of newly green buildings with the other existing green office buildings in Malaysia.