The main purpose of this paper is to present the experimental investigation on the axial behaviour of concrete-filled steel tube columns using semi-lightweight aggregate concrete as an infill type. The paper describes the effect of variation in the diameter (D), width (B), thickness (t) and length-to-diameter (L/D) ratio of steel tube with a special type of core concrete on the ultimate axial load (N-ue) and the associated axial shortening (delta(ue)) for both the circular and the square CFST columns. The experimental campaign involves, finding the mechanical properties of steel, concrete and the Axial compression tests conducted on the CFST specimens. The 2(k) factorial design; a DOE approach is implemented primarily. Subsequently ANOVA, a statistical tool is used to analyze and compare the different combinations for reducing the error variance and finding the most influential factor on N-ue and delta(ue), Basic prediction models are proposed to predict the response N-ue and delta(ue) for the CFST columns and further, the Response surface method is implemented to revise the models for interaction between the different combinations. Comparisons are made with the existing design codes like ANSI/AISC 360-10 and EN-1994 (Eurocode 4) to validate. The proposed model predictions thus, validate the experimental results with good accuracy. The tests confirmed the considerable influence of semi-lightweight aggregate concrete on the performance of CFST columns. In addition, the confinement improves the strength of the infill concrete and thus validates the use of semi-lightweight aggregate concrete as an infill to the CFST columns for structural purpose and particularly for reduced self-weight, (C) 2019 Elsevier Ltd. All rights reserved.