You can find the value of pKb of your molecule using a thermodynamical cycle (thus necessiting more than one calculation).
In the past, I managed to predict with a good accurary pKa of different molecules when linked to a metal centre. I good suggest you to have a look on this tutorial online:
You would have just to adapt your thermodynamic cycle to have the value of pKb instead of pKa or simply evaluate it via the relation with pKe in water.
The easiest way to compute a pKa value in Gaussian is to use the isodesmic method. This allows you to avoid any complex thermodynamic cycles. In this method you use a reference with known pKa (e.g. formic acid) to compute the pKa of interest. Please be also aware that the implicit solvation models implemented in Gaussian are far from perfect. Fortunately it fails in a predictable manner so you can scale the obtained raw pKa value in order to obtain a reasonable guess. Using this method you can get pKa values with an accuracy of roughly 1 pKa unit with minor computational costs.
A detailed description of the method, its accuracy and suitable scaling relations can be found here: