In my opinion it depends on your case study, It is better to use different indices for almost 30 past years and with other evidences then you can make decision which on can work for your case study.
As you know, if we compute just one index such as SPI which uses just precipitation it wouldn't really perfect, but if you calculate PDSI which needs several input data such as field capacity, soil moisture and evapotranspiration then you can get a result with minimum uncertainty.
It is not really easy, but as I mentioned before, in a glance by comparing different results of the indices for the specific case study, you can make a reasonable decision.
In this regard, I want to inform you an important tool which you can easily calculate around 20 indices without any concerns. See the bellow link for further details, I think it works for you as an applicable tool.