Composite materials are a common choice for designing light weight aeronautical structures. The design of these structures consists in defining the stacking sequence of each zone such that it the weight of the structure is reduced and the mechanical reserve factors are preserved. \\ The Manufacturers impose some design rules to be applied to the stacking sequences. These rules aim at preventing from some failure modes like delamination or buckling. Moreover, some rules related to the thicknesses of the zones which are derived from engineering practice have to be satisfied.\\ The mechanical responses are computed using numerical simulations. Therefore, the design and optimisation of a structure can require a large number of numerical simulations and this can be costly in computations. Therefore, the mechanical responses are approximated using surrogate models based on a Design of Experiments (DoE) with limited number of numerical simulations. The configuration of the stacking sequences of each point in the DoE must satisfy the prescribed design and manufacturing rules.\\ In this work, a sampling method is presented to generate a DoE with constrained samples. The set of constraints are the design and manufacturing rules. Given that the stacking sequences are parametrised using integer numbers, a backtracking procedure is used to generate admissible points. The quality of the surrogate model based on the DoE is insured by a space-filling criterion. The points of the DoE are generated by maximizing the minimal distance between them. This can guarantee a good coverage of the design space and it can integrate all the different mechanical behaviors of the structure.\\ The sampling method is implemented using a constrained simulated annealing algorithm. Three test cases taken from industrial applications are presented to show the efficiency of the proposed method.