In recent years, high-entropy alloys (HEAs) are attracting significant attention owing to their distinctive design adaptability and exceptional properties. Herein, machine learning methods namely extra tree (ET), K-nearest neighbors (KNN), random forest (RF), support vector regressor, and linear regression are utilized to predict the mechanical properties of MoNbTaTiVAlx refractory HEAs across varying compositions and temperatures. By doing so, the study aims to minimize the dependence on experimental testing. Among the models, ET, RF, and KNN exhibit superior predictive performance, achieving R2 values of 0.998 which closely align with experimental results. Additionally, a new stress-strain curve is generated for an aluminium composition of 0.4, with the ET, RF, and KNN models maintaining high predictive accuracy with R2 values of 0.985, 0.978, and 0.97, respectively. This innovative application of machine learning significantly reduces the need for exhaustive experimental testing, resulting in considerable savings in resources and accelerating advancements in HEA research and development.