The full factorial design gives the best results, because evaluates responses for all combinations of design variable levels, but it is costly(computation time and resources) for large number of design variables and levels.
You can download this eBook(https://altairuniversity.com/free-ebook-design-of-experiments-with-hyperstudy-a-study-guide/) for more information about the DOE and DOE types.
It is difficult to answer the question, as it depends entirely on the problem. What do you want to do? maybe it helps if you realize what is the difference between that two designs. While factorial designs are used to model linear effects, center composite designs have the goal of modelling non-linear effects. The accuracy increases naturally with the number of setpoints. CCD usually has more setpoints and therefore provides higher accuracy. But I always recommend to start with a fractional or full factorial design with central points and then expand to a CCD if necessary.