Mining data securely and effectively from multiple sources has become a critical issue in the era of rapid information growth. This paper thus addresses the federated utility mining problem within a horizontal framework, in which items are nearly identical across clients, but users differ. We propose a federated utility-mining algorithm that transmits local high-utility itemsets, rather than transaction data, from clients to a server, thereby increasing data privacy. Clients independently mine high-utility itemsets, and the server integrates these itemsets to produce final results. Although the proposed algorithm discovers fewer high-utility itemsets than the traditional utility mining algorithm, it can save significant integration time and avoid transmitting the original datasets. We then conduct experiments on two real datasets by allocating the data among three clients, with the results showing that the proposed algorithm achieves good performance.