As the new round of technological revolution and industrial change evolves, data has become a valuable strategic resource and a new factor of production. The focus of the discussion on data science has shifted from "whether data can be used" to "how to leverage the positive value of data". At home and abroad, some typical enterprises have emerged with data for good, such as DXY and Baidu Maps, which have created great business and social values. However, more enterprises are at a loss of what to do in the face of data for good, and some of them have even made data wrongdoings such as data killing, illegal data collection and sensitive data leakage. The reason for this is the lagging awareness of data for good in the practical world. However, data for good research is in its infancy, and only basic issues such as definitions, dimensions, and scales have been initially explored. The lack of influential research results on the key issue of how data for good creates business and social value has led to the limited guidance of existing research on the practice of data usage. Therefore, it is of great theoretical and practical significance to explore the mechanism of data for good affects value creation.In fact, there is a natural link between data for good and digital innovation capability. As data is the core of digital resources, enterprises need to build and cultivate their innovation capability to creatively utilize, coordinate and develop digital resources at different levels and areas of their business and operations, which is an important prerequisite for value creation. It can be inferred that synergy between data for good and digital innovation capability is the key to value creation. Specifically, this paper focused on the following three questions: what are the antecedents of data for good? how can the synergy between data for good and digital innovation capability create value? And how is the path of value creation realized?This paper took DXY and Baidu Maps as examples, and used the grounded theory research method to answer the above questions. A "motivation-behavior/capability-output" mechanism model was constructed for the synergy between data for good and digital innovation capability to create value. The research found that the motivations for data for good include external pressure and incentive factors from the market, law, and multi-actors, as well as internal subjective and objective factors in terms of strategy, spirit, and complementary assets; the synergy between data for good and digital innovation capability affects value creation, and the mechanism is a dynamic enhancement process in which the two influence each other and promote each other; there are two main paths for data for good to create value. Firstly, based on external pressure and incentive factors, enterprise should make a responsible use of data and realize business value in the process of interacting with digital agility capability. Secondly, based on internal subjective and objective factors, enterprise should conduct data charity, and realize social value in the process of interacting with recombinative capability.This paper further clarified the motivation of data for good, the mechanism path of value creation through the synergy of data for good and digital innovation capability. Therefore, the theoretical contributions of this paper are as follows: First, this paper discussed the motivation of data for good, and responded to the call of scholars to discuss the antecedents of it, which is of great significance to enrich the theoretical study of data for good. Second, this paper clarified the synergistic effect between data for good and digital innovation capability, expanded the theoretical research on the interaction mechanism, antecedent, and results, as well as research methods of the two, and promoted the establishment of a deep connection between data ethics research and digital capability research. Finally, this paper proposed a mechanism model for the synergy between data for good and digital innovation capability to create value, which will not only provide a clear research framework for following studies, but also help to reveal the laws and essence of data for good more clearly. More importantly, there is still a long way to go in the development of data for good research and practice, and this paper is only a small step to further enrich the data for good research in China and is hoped to provide wisdom for the construction of theory on data for good.