With the "post-processing" technique we explore the role of the convective overshooting on the production of s-nuclei in stellar models of different initial mass and metallicity ($15 \leq M_{ZAMS}/M_{\odot} \leq 25$; $10^{-4} \leq Z \leq 0.02$), considering a range of values for the parameter $f$, which determines the overall efficiency of convective overshooting.We find enhancements in the production of s-nuclei until a factor $\sim 6$ (measured as the average overproduction factor of the 6 s-only nuclear species with $60\lesssim A\lesssim90$) in all our models of different initial mass and metallicity with $f$ in the range $0.01{-}0.035$ (i.e. models with overshooting) compared to the production obtained with "no-overshooting" models (i.e. models with the same initial mass and metallicity, but $f=10^{-5}$). Moreover the results indicate that the link between the overshooting parameter $f$ and the s-process efficiency is essentially monotonic in all our models of different initial mass and metallicity. Also evident is the higher s-process efficiency when we progressively increase for a given f value both the mass of the models from 15 M$_\odot$ to 25 M$_\odot$ and the Z value from 10$^{-4}$ to 0.02. We also briefly discuss the possible consequences of these results for some open questions linked to the s-process weak component efficiency, as well as a "rule of thumb" to evaluate the impact of the convective overshooting on the yields of a generation of stars.