Homology modeling will try to model the regions of your protein that are similar in sequence to other proteins of known structure in a given database. It is possible that certain regions of your protein will not be modeled. That being said you can check at: http://modbase.compbio.ucsf.edu/modbase-cgi/index.cgi
if your protein is not there try using http://swissmodel.expasy.org/ This tool will get your query protein sequence and try to find templates that exist in their database.
There are also other automated servers such as I-TASSER
You can either use native I-TASSER algorithm which takes into account existing templates or the QUARK which does ab initio prediction, meaning from scratch, without templates and only taking into account the amino acid sequence.
You can also try the automated PHYRE2 server
Here you will have 2 modes, normal and intensive. Templates are used here as well. For all residues you will get a key of confidence scores giving you an idea about the reliability of the model/region that is modeled.
All these are some guidelines. The aforementioned websites will send you results within 1-4 days, no additional work needed other than submitting your fasta sequence. The resulting models should be interpreted carefully, trying to take into account proteins with similar function/structure with known 3D confimation where possible,
I hope this helps,
Good luck with your research,
Further to Ibrahim's reply:
If your protein sequence is in UniProtKB, you can have a look at the "Structure" section of the entry, which links to various model databases, even if no cross-references to PDB are available, e.g. ProteinModelPortal, SMR (Swiss-Model repository), ModBase and MobiDB, e.g. http://www.uniprot.org/uniprot/Q96J66#structure