I'm working with the aminoacid sequence of a capsid protein of a potyvirus and I want to know which method in Mega 7 is appropiate for a phylogenetic analysis
Selection of algorithm , depends upon the similarity of the sequences which you are using. I guess you have a set of sequences to begin with . If extent of similarity is high then choose Maximum Parsimony, for moderate similarity choose Neighbor Joining and in case of low similarity , best option is Maximum likelihood. Maximum likelihood is very exhaustive method and provides good result. Further more you can perform Molecular Clock Analysis as well.
The more adeccuate is to select the apropiate model according your data. For this, you must in first place to align your sequences and save as mega format. Very important too is to select an adecuate outgroup for your phylogenetic reconstruction. You must select Analysis->Models->Find best DNA/protein models and load your protein alignment. With this data you can select the most correct model according yor data and use it for the phylogenetic analysis.