The primary purposes of this paper are to conduct a sensitivity analysis (SA) of the square composite multi-layered slab and to optimize the maximum fundamental frequency. The analysed composite slab consists of eight layers, in which the angles of lamination in the particular layers are assumed to be symmetrically established, relative to the middle surface of the slab: [λ1/λ2/λ3/λ4]s. The lamination angles are the project variables to be established during the maximization of the first natural frequency (see [1]). During the analysis, the first natural frequency is converted into the non-dimensional frequency parameter Ω (see [2]). Eleven different variants of boundary conditions are considered. The analysis begins by building the Finite Element Method (FEM) model of the slab. The model was prepared with the ANSYS PyMAPDL library. Then, the patterns for all 11 variants of the boundary conditions for the purpose of developing the deep network surrogate model are generated in the Python environment. The next part of the analysis is the sensitive analysis, which means determining which angle: λ1, λ2, λ3, λ4, affects on the value of Ω in the most part. In these analyses, the Morris method was applied. The trained surrogate model is also utilised to calculate the objective function values in the genetic algorithm (GA) based optimization. The set of optimal lamination angles and the value of the Ω parameter corresponding to the optimal set of angles are the results of the computations from the GA. The main advantage of this approach (SA + DNN + GA) is a very short duration of the sensitivity analysis (less than 1 second) and the optimization procedure. Funding: Financed by the Minister of Science and Higher Education Republic of Poland within the program “Regional Excellence Initiative”.