In this paper, a regression model is developed to predict and optimise the compressive strength of periwinkle shell aggregate concrete using Scheffe’s regression theory. The results obtained from the derived regression model agreed favourably with the experimental data. The model was tested for adequacy using a student t-test at 95% confidence level and was found to be adequate. A computer programme coded in basic language was used to select the mix ratios that optimized the compressive strength of periwinkle shell aggregate concrete. The optimum compressive strength was found to be 19.50N/mm2corresponding to a mix ratio of 1:3:6 (cement, sand and periwinkle shell) at a water-cement ratio of 0.65. With the formulated model, the mix ratios corresponding to a desired strength value can be predicted with reasonable accuracy and without waste of time. http://dx.doi.org/10.4314/njt.v36i1.5