Usually RMSD below 2Å is considered good for Homology modeling study. However the value you are geeting is not much high from standard value therefore you can consider your model for futher study. Aditionally you can validate your model using other methods such as steriochemical properties analysis using PROCHECK, ProSA and Verify 3D. Also short time MD simulation can be done to refine the model. Read some papers based on Homology modeling for more clarifications.

In addition to what Vikas Kumar said, I would suggest you to try SAVES (https://saves.mbi.ucla.edu), where you can run different programs to assess your structure.

In my opinion, the RMSD you should be analyzed with other information you have about your protein. How much identity does it have to other known structures? This RMSD is compared to what? Where did you get your homology model?

This should help you: https://www.umass.edu/molvis/workshop/homolmod.htm

Best,

Lorenzo

There is no threshold for "acceptable" RMSD.

in protein modeling with homology modeling, RMSD is not a consideration for choosing a model. other parameters can be considered: for example: sequence coverage, sequence similarity, confidence number. Furthermore, the obtained model must be evaluated with certain parameters, such as the Ramachandran plot and others. For the validation process, the SAVE v.06 server can be used. Refining the model can also be done with the server refinery tool.