Drilling operations generate CO2 emissions from diesel fuel used to power generators and other equipment on drilling rigs as extraction activities create around 10-15% of total energy-related emissions globally, an equivalent of 5 billion tonnes of greenhouse gas emissions. Improving operational efficiency is crucial for reducing overall emissions and combating climate change. Therefore, recognizing more efficient approaches and solutions with reduced carbon footprints is essential in creating future business continuity. This paper presents the application of machine learning optimization techniques in carbon emission management through technical limit performance benchmark. A Machine Learning (ML) model utilizing basic drilling parameters such as weight on bit (WOB), rotary speed (RPM), penetration rate (ROP), and torque (TOR) was used to estimate instantaneous Unconfined Compressive Strength (UCS) within the technical limit of operational performance across the different Lithologies. Supervised learning algorithm with 70:30 data splits using four ML algorithms, including Random Forest, support vector regression, artificial neural network (ANN), and Categories boost, was utilized in the prediction. The estimation was then leveraged to determine the maximum achievable penetration rate baseline for various lithologies based on rock drillability. Enabling a comparative assessment between actual drilling performance against an engineered technical limit drill rate performance as well as the expected greenhouse gas emissions (GHG). A case study using a field dataset showed that the models predicted instantaneous UCS based on the low- cost drilling parameters other than using the high-cost well-log and core sample techniques. Prediction accuracy was within the acceptable margin of error with CatBoost and ANN as best predictors, with correlation coefficient (R2 ) values of 0.75 and 0.77, respectively. The defined baseline performance benchmark improved drilling performance by 30-60% with a reduction in CO2 emissions across the project by 20 tCO2e emissions, equivalent to a 50% reduction in carbon emissions. The result of this research is essential for the continuous improvement in drilling carbon emission, performance monitoring, and improvement of overall drilling performance.