The genetic algorithm (GA) model presented here provides specific JONSWAP parameters that can be used for wave modelling. This work describes a validated heuristic model based on GA, to select JONSWAP spectra parameters, regardless of water depth restrictions and sea state conditions. The identification of the JONSWAP spectra parameters is difficult, as the alpha and gamma coefficients have scattered distributions that modulate the spectral peak energy. In addition, the selection of alpha and gamma coefficients from in situ free surface records may be difficult and time-consuming, because of the amount of data and nonlinearities involved. The proposed model uses either in situ or numerically modelled wave data and has three main steps: (1) generation and crossover, (2) minimisation of the cost function ΔHs, defined as the minimum difference between the calculated artificial significant wave height and the in situ wave height (instrumented or modelled), and (3) mutation and natural selection. To apply the model, in situ wave data measured by an acoustic Doppler current profiler over 5.5 months was used in this research. The results show a high correlation (r2), of 0.95, between the best fitted curves of modelled spectra and measured data.