Plasmonic metastructures represent a class of bidimensional metamaterials that can be designed and optimized to realize optical sensors aiming a wide variety of actual applications. In this work we propose a modelling approach for such metasurfaces based on a customized particle swarm optimization (PSO) algorithm implemented in the commercial Ansys Lumerical FDTD software. Providing to the algorithm the relevant optical and morphological parameters, we obtain a set of optimized geometrical parameters tailored to the required usage based on a specific fitness function mimicking the actual experimental measurement configuration. The requirements can be associated with different lithographic techniques. Specific constraints on piloting compatibility and scaling up processes are also implemented. Two different array dispositions have been analyzed: square and hexagonal arrays of air cylinders in a gold layer on a glass substrate. In our more recent activities, we are dealing with gold nanohole arrays for biosensing applications for biosensors based on surface plasmon resonance (SPR) measurements. Consequently, we focused our study on tuning the main localized plasmon resonance in the near infrared to maximize the effectiveness in SPR detection. In the end, two suitable structures have been successfully defined, one for each array disposition, that we have already planned to fabricate, in order to validate our design routine.