Data-driven material discovery has recently become popular in the field of next-generation secondary batteries. However, it is important to obtain large, high quality data sets to apply data-driven methods such as evolutionary algorithms or Bayesian optimization. Combinatorial high-throughput techniques are an effective approach to obtaining large data sets together with reliable quality. In the present study, we developed a combinatorial high-throughput system (HTS) with a throughput of 1000 samples/day. The aim was to identify suitable combinations of additives to improve the performance of lithium metal electrodes for use in lithium batteries. Based on the high-throughput screening integrated with data scientific technique, over 10,000 samples, a specific combination of five additives was selected that drastically improved the coulombic efficiency (CE) of a lithium metal electrode. Importantly, the CE was remarkably decreased merely by removing one of these components, highlighting the synergistic basis of this mixture. The results of this study show that the HTS presented herein is a viable means of accelerating the discovery of ideal yet complex electrolytes with multiple components that are very difficult to identify via conventional bottom-up approach.