Below is a good, brief paper that talks about the pros and cons of different RAD-seq methods. I would suggest reading it before you make any decisions.
I think is best to think about what technique is best for your lab (so not just this project but also the next) since there is an upfront cost to all the methods. My personal preference is ddRAD since it is probably the most flexible and thus good for labs with a diverse array of projects. However, if your lab specializes in any particular area another method may be better. Hope this helps.
2bRAD is best (of course I should say that). No secondary digestions, no intermediate purifications, no size-separation in gels. Just add reagents consecutively to a well in 96-well plate, amplify, pluck a distinct band from a gel, done. Bioinformatics is good, too: https://github.com/z0on.
Here is a 2bRAD-based linkage mapping project, just published in Science:
Both ddRAD and GBS could offer you hundreds of SNPs, it's hard to say which is better. if you have a reference genome with you species, low coverage GBS sequencing is recommended for the low cost