The problem now I will solve is that: we have 3 species, such as A, B, C. All of them are primates and have reference genomes. And we hope to get highly accurate multiple alignments of the whole genomes for further study.
If the three primate genomes you are interested in are in the list of publicly available primate genomes in UCSC/Ensembl/NCBI, you could just use the pre-calculated multiple whole-genome alignments in Ensembl Comparative Genomics:
If the species you are looking for are incomplete or not included, or you'd like to do the alignment independently with different parameters, methodology, etc, Multiz is a good start. Vista can be used to add sequences too.
I've used Mauve (http://gel.ahabs.wisc.edu/mauve/) in the past. If you're aligning relatively complete assemblies, it might be what you're looking for. It also has a feature ("move contigs" or contig reordering) where it will rearrange contigs (from, say, a draft assembly) until they're in an order / orientations that maximize alignment with a trusted reference.