Nowadays consumer loan plays an important role in promoting the economic\ngrowth, and credit cards are the most popular consumer loan. One of the most\nessential parts in credit cards is the credit limit management. Traditionally,\ncredit limits are adjusted based on limited heuristic strategies, which are\ndeveloped by experienced professionals. In this paper, we present a data-driven\napproach to manage the credit limit intelligently. Firstly, a conditional\nindependence testing is conducted to acquire the data for building models.\nBased on these testing data, a response model is then built to measure the\nheterogeneous treatment effect of increasing credit limits (i.e. treatments)\nfor different customers, who are depicted by several control variables (i.e.\nfeatures). In order to incorporate the diminishing marginal effect, a carefully\nselected log transformation is introduced to the treatment variable. Moreover,\nthe model's capability can be further enhanced by applying a non-linear\ntransformation on features via GBDT encoding. Finally, a well-designed metric\nis proposed to properly measure the performances of compared methods. The\nexperimental results demonstrate the effectiveness of the proposed approach.\n