The first objective of this study was to develop a model for the peak concrete temperature (T) based on predictors known as its primary influencers and verified by an artificial neural network (ANN). The second objective was to gauge the influence of the level of replacement of Portland cement (PC) with fly ash class F (FA) and slag cement (Slag) and the variation in the cementitious materials content in concrete (Wc) in typical mass concrete mixtures on T. The least dimension (LD), the concrete temperature at placement (CT), Wc, Slag and FA contents, in addition to tricalcium silicate (C3S) and tricalcium aluminate (C3A) in PC and the Fineness (F) of PC, were investigated. The temperatures of 111 massive elements with FA were simulated 43,768 times with the Levenberg-Marquardt (LMA) and Bayesian (Bayes) models in the ANN. The temperatures of 72 massive elements with Slag were simulated 78,255 times. The T of massive elements appeared better characterized by multivariate regressions. The variables that significantly affected the T seemed to be the LD, CT, and C3S content in PC-FA systems and the LD, CT, and Slag in PC-Slag systems. The LD appeared to exercise the most significant influence on T in both cementitious systems. High substitution levels of PC with FA and Slag should generate different effects. Slag appears to be more significant to T in this scenario. The accuracy of multivariate models trained by ANN algorithms and the low significance of Wc to control temperatures of a particular concrete class were revealing.