The problem of poverty always gets the main attention and becomes a fundamental problem in Indonesia. Therefore, the government as the most responsible party needs to overcome this problem. But the problem of poverty is not only the government's responsibility, but it also requires the participation of the community. Therefore non-governmental organizations or institutions such as Baitul Maal or Amil Zakat institutions in the regions must be able to optimize their role in the poverty problem. One effort from Amil zakat institutions in contributing to solve the problem of poverty was increasing the loyalty of donors. Because the number of donors which have been registered and the higher of donor's loyalty is one of the benchmarks of success, it will be directly proportional to the donation collection funds that will be obtained. But in fact, to increase the number and loyalty of donors is not easy, this is due to the determination of marketing strategies or marketing maps which were conducted so far was not optimal. Therefore in this research, a marketing strategy had been prepared or marketing maps had been conducted using the data mining approach. The first method that had been used as K-Means Clustering was useful for grouping donor data. Donor data that will be processed is transaction data for the last 6 months at an Amil zakat institution in the Karawang area with the RFM (Recency, Frequency Monetary) attribute had been selected. Then from the search results which using the elbow method, the most optimal value of k or number of clusters was 2 and produced a low donor group and a high donor group. And in the final process, an association process was carried out on the transaction data from each group of donors by using the Association Rules method to determine the relevance of the donation programs that had been chosen by each group of donors. Then, the results can be used as data or references in arranging marketing strategies maps.