Abstract : Geo-polymer composites are emerging as sustainable replacement options of conventional cement concrete, for effective reduction in carbon footprint. Owing to the use of supplementary cementitious materials viz. fly ash, GGBS, metakaolin, brick husk ash and their combinations, the resulting concrete is different from the conventional concrete in terms of fineness, density, packing etc. Hence, the correlation equations used to derive compressive strength values from non- destructive test results, for conventional concrete, cannot be applied to the Geo-polymer concrete. In the recent experimental research program, six types of geo-polymer concrete specimen were prepared to evaluate their performance. Both nondestructive tests (Schmidt rebound hammer, Ultrasonic pulse velocity) and destructive tests were conducted to determine compressive strength using suitable regression equations. The collected data from tests were first imported into predefined models for analysis of polynomial regression technique (PRT) to derive the equations. To improvise the data structure performance and avoid issues such as over fitting and under fitting, the datasets were imported in machine learning algorithms, including ANN, ANFIS techniques. If we compare regression analysis, ANN and ANFIS, ANN the latter yields less error than the former.