In a few words: EGA is previous to everything, though one can use SGA 1 to complement some aspects of EGA IV. FGA goes "in between". There is a complicated tree for SGA. SGA 3 is independent of the rest while SGA 4, SGA 5 and SGA 7 are a full saga. SGA 6 is more or less independent but you need Verdier's thesis (or its resume at the end of SGA 4 1/2). In fact, SGA 4 1/2 can be considered as an introduction of certain aspects of SGA 4. FGA has three topics: formal schemes, duality and representable functors and should be read "as needed". The new edition of EGA I is a very nice reference.
It is perhaps interesting to point out that it is unrealistic try to master all of this material in a short amount of time. Perhaps one should study one of the manuals and then rely on EGA and SGA for further topics and additional details.
To mention a few good references (without being exhaustive): Hartshorne, Götz-Wedhorn, Mumford-Oda, Liu and Bosch.