I have a EM map (.mrc), with fitting and based on a crystal structure I got the fitted PDB file. I convert the mrc map to the CCP4 map，

no need to convert .mrc into .ccp4 (isn't it the same thing?). Anyways, phenix.real_space_refine can use both.

and I run phenix.real_space_refine model.pdb map.mtz run=adp, with phenix-dev-2034, in order to get my own b-factors for my fitted PDB based on my EM map. However after the above phenix.real_space_refine, I checked the refined PDB, and I find phenix.real_space_refine neither change the coordinates, nor change the b-factors for each atom.

I need to see the files to be able to comment. If you share the files (off list and not sending directly via email but using things like Dropbox instead) then I will be happy to have a look.

Thus will you please introduce to me how can I get my own b-factors for my fitted PDB in order to deposit my map and fitte pdb to the EM database?

phenix.real_space_refine model.pdb map.mtz run=adp should just do what you want. If it didn't - see above (I need to see the files). Pavel