In this paper we introduce a new packing problem, the Three-Dimensional Knapsack Problem with Balancing Constraints (3BKP), the extension of the standard Three-Dimensional Knapsack Problem (3KP) where additional constraints related to the center of mass of the three-dimensional packing are given. Given a set of box items i = 1,. .. , n with sizes w i , d i and k i , a profit p i , and a mass m i and a container called knapsack of fixed dimensions W , D and H, 3BKP consists in orthogonally packing a subset of the items into the knapsack in order to maximize the sum of the profits of the loaded items. The items must be accommodated into the knapsack such that they do not overlap. Moreover, the center of mass of the overall packing must lie into a predefined boxed domain within the knapsack. We assume that items can be rotated. We give a MIP formulation of the problem, used to derive bounds, as well as an efficient heuristic method able to solve, with a limited computational effort, the test instances. Moreover, new test instances are introduced and used to derive extensive computational results. The results show how the MIP model is able to find better bounds than other relaxations, and how the heuristic method is able to efficiently solve both instances explicitly designed for 3BKP, as well as to be competitive with methods explicitly designed to solve 3KP.