<p>Nowadays, one of the most often utilized concrete types is high performance concrete (HPC). A variety of cementitious ingredients, including fly ash, metakaolin, and silica fume, are commonly added to increase the material’s durability and compressive strength (CS). Time and money can be saved by giving designers and structural engineer’s access to trustworthy and precise models for calculating CS. As a result, trial patches of HPC in site will be reduced due to the conclusions of this study. Appropriate component selection and proportioning are necessary for the production of high performance concrete, which is primarily distinguished by its fine pore structure and low porosity. Therefore, genetic algorithm (GA) and swarm particle optimization (SPO) were applied to existing two groups of HPC: (40–80) MPa and (80–120) MPa. Each group contains certain number of training mixture and verification sets. The results were compared with the outputs of both of GA and SPO. It was found that there is significant correlations between experimental results and GA and PSO. Generally, the errors were minimized with GA and PSO with both of groups with different variables. The average error was higher in the ingredients of sand/total aggregates ratio (s/a) due to it was no concern to the size and finesse modulus which affects directly the CS. Maximum deviations were that of chemical admixtures in two groups due to the type and content and unexceptional conditions in mixing process. Minimum deviations were with the water content with both of GA and SPO in two groups.</p>