This current research describes and provides a guideline on the procedure for mixed proportions of self-compacting concrete (SCC). Experimental design, response surface methodology, contours, and surface response were used for model prediction and optimization of responses of the SCC parameters. The significance of this research is focusing on constituent materials and their corresponding proportions in SCC. Furthermore, there is still a lack of standardization in the design of SCC mixes. In this paper, the effects of constituent material parameters on fresh and hardened properties of SCC were explored using response surface methodology and Artificial Intelligence. Different compositions of SCC were designed based on cement, sand, coarse aggregate, fly ash, and superplasticizer. The J-ring test was used to study the fresh state characteristics, and compressive strength was measured at 28 days. Experimental results were modeled using multiple regression to predict the effect of input variables on performance responses. The fitness of the models was also validated by an analysis of variance ANOVA analysis, which supported that a complete quadratic model successfully described the combined effects of components on SCC. In addition, artificial neural networks (ANNs) were utilized for predicting and optimizing SCC characteristics. The actually ANN-based predicted values presented better accuracy than that of the conventional statistical models, which had great value in mix design optimization. In general, the study demonstrates that integrated use of RSM and ANN will provide a powerful and effective tool for modeling, prediction, and optimization for both fresh and hardened properties of SCC.