An artificial neural network aided methodology succeeded to improve the catalytic performance of Co/SrCO3 for dry reforming of methane at 1 MPa, 1023 K, SV = 100 000 mL/h/g, feed composition = CH4/ CO2/N-2 = 45/45/10. CH4 conversion after 100 h time on stream was enhanced from 18% with Co/SrCO3 to 33% with the optimum Co/SrCO3, and CH4 conversion was constant from 10 to 100 h with the optimum catalyst, whereas it decreased gradually with Co/SrCO3. Reduction of carbon deposition after 100 h run (from 18.6% with Co/SrCO3 to 5.8% with the optimum) contributes to the stability of the optimum catalyst. In the first step of the optimization, new additives to the Co/SrCO3 catalyst were screened by using an artificial neural network (ANN). Catalytic activities of 10 mol % CO + 1 mol % X/Sr-3 (X = B, K, Sc, Mn, Zn, Nb, Ag, Nd, Re, or T1) and 16 physicochernical properties of these 10 elements were used as training data of the ANN, and the effect of each additive among 53 elements was predicted by the trained ANN. From the prediction and experimental verification, no better additive than the training data was found. Hence, high CO yield was obtained by Re addition, and CO yield gradually increased by Nd addition. Then, the loading of Co, Re, and Nd was optimized by means of design of experiment (DOE) and ANN for higher performance in the second stage of the optimization. The optimum catalyst was decided as 4.3 mol % Co, 2.2 mol % Nd, and 2.3 mol % Re supported on SrCO3, and it showed the remarkable performance.