The happiness maximization query which overcomes the inherent drawbacks of the top-k and skyline queries has been an important tool for multi-criteria decision making in the last decades. To ensure decisions made by these queries to capture users’ diverse needs and preferences, and to be more applicable against data discrimination and bias, fairness from the data-side and user-side needs to be considered simultaneously. In this paper, we consider the problem of answering the happiness maximization query with two-sided fairness which has never been investigated before. By truncating the happiness ratio function of the query, we transform our problem to a submodular function maximization problem with constraints, and the data-side fairness is modeled similar to the matroid constraint while the user-side fairness along with the happiness maximization query is treated as a multi-objective optimization problem. Further, we provide the Greedy2F algorithm by incorporating the elaborate techniques of above two areas. Extensive experiments on both real-world and synthetic datasets illustrate the effectiveness and efficiency of our proposed algorithm.