By Taguchi design method-L_9(3^4), the parametric combination for maximum retardance in citrate-coated ferrofluids (FFs) is determined. Multiple regression (MR) is completed by Excel. Moreover, Taguchi measured results were used as training data in the feed forward artificial neural network (ANN). Computed results in MATLAB showed the ANN model had a high correlation coefficient of 0.9999998, close to 0.9999999 obtained by MR. Hence the trained ANN was able to predict the maximum retardance of 33.41°, with a relative error of 2.17% to MR.