The use of recycled concrete aggregate obtained from construction and demolition wastes in concrete production is gradually becoming a sustainable approach to curb the incessant challenges linked to resource conservation in the construction industry. However, much is yet to be understood on how the properties of recycled aggregates could influence the mechanical strength. In situations where there is evidence of the consequential effect not much had been done in utilising machine learning algorithms for modelling the effect of the process parameters on the compressive strength of recycled concrete aggregates. Therefore, this study utilised fourteen (14) input parameters were as predictors for determining the compressive strength of the concrete up till 90 days of curing. Feature ranking, hyperparameter tuning, and model evaluation were implemented to better understand the predictive accuracy of the machine learning models adopted. The outcome of the feature ranking revealed that the cement content, superplasticizer dosage, and the size of recycled aggregates as paramount parameters in predicting the compressive strength of RCA concrete, with corresponding importance scores of 48.4, 47.08, and 42.81, respectively, based on F-test. The adoption of Gaussian Process Regression Exponential models emerges as an optimal choice based on their high R2 for training, testing, and validation (0.92, 0.89, 0.98) for enhancing predictive accuracy of recycled aggregate concrete, thereby contributing to more efficient decision-making processes in the field of concrete technology. Ultimately, the collective performance of these models signifies a consistent ability to achieve reliable predictions with minor discrepancies, emphasising their applicability and utility in real-world scenarios.