Sure you can import the complex into Maestro. After that you must run protein preparation wizard and the obtained complex is now suitable for desmond calculation.
In principle, yes you can directly use the docking results as input for Desmond MD. Beware, however, that the PDBQT files used in VINA lack the non-polar hydrogens and usually carry Gasteiger charges.
To accomplish your task, you will need to:
1. convert the input files from PDBQT format to some Maestro readable format (should be fairly straightforward)
2. merge the ligand and protein files (easy within Maestro)
3. Add missing hydrogens, paying due attention to potential changes in the protonation state of your ligand, which is easily achievable using the Protein Preparation Wizard as Simone Brogi suggested
4. I would recommend doing a short minimization under the forcefield you intend to use for the MD before starting the MD simulation to relieve minor strains in the complex (default relaxation protocol of Desmond can mostly take care of such non-optimal geometry)
Perhaps Maestro can directly read PDBQT format, I would give it a try first.
But in any case, the best way in my opinion is to use AutoDockTools (http://autodock.scripps.edu/resources/adt) to convert ligand and protein PDBQTs into some other Maestro-supported format (e.g. mol, mol2, or pdb).
Other options include: using openbabel (not very reliable for PDBQT format from my brief experience) or opening the PDBQT file in PyMol and re-saving it in a different format.