When constructing a phylogeny, the main thing to look for is that the sequences you are aligning are homologous. When you say gene cluster, do you mean a multi-copy gene family? If all the genes in the cluster are orthologs/paralogs (with multiple copies derived via gene duplication, etc) then you will be able to align them all together and get a phylogeny. The only issue would be if the genes in the cluster are so deeply diverged that you can no longer accurately align the sequences. If that is the case, you would be better off to divide them into separate orthogroups that you can align.

There are a lot of different phylogenetic methods, but a good one to start with is maximum likelihood, and the program IQtree is a pretty straightforward way to make a maximum likelihood tree.