Hello, Would you like to elaborate you question a bit more. "Molecule", did you mean a protein? Do you have any interface data for those two molecules?
From the technical point of view, it's as simple as merging two text files with coordinates, if you have them in an text format, like xyz (just need to change the first line with the total number of atoms) or pdb (unless you have important CONECT lines). Most visualization software have also the merge option, I particularly like the one from shcrodingers maestro, but any will do. If I remember correctly pymol can do the trick very well.
That's obviously possible of you know what is the orientation of the molecules with respect to each other. If you don't, then it becomes a non-trivial task.
For small molecules: you can build manually all possible orientations and optimize all of them (perhaps using a fast, semiempirical method like PM6, which works well for organic systems). The one with the lowest energy is the probable geometry of the complex. DFT optimization would be better, but it may become expensive if you have many starting orientations. There are also software for performing automatic conformational search, but I'm not sure if they can handle complexes (they can for sure handle single molecules).
For proteins / biomolecules: there are freely available servers for protein-protein (also for protein-DNA/RNA) docking, semi-reliable and only work well if you have some experimental information about the interface and can narrow down the interacting residues (and also get any validation).
Through Hyper-Chem it is possible to merge your system. Open a new work-space and build your macro molecule( as example a strand of DNA) . Now merge the macro molecule (such as DNA) and your molecule so that they are both on the screen at the same time. You will use the molecular mechanic method AMBER while testing the interactions of your molecule with the DNA.
Thanks for your answers. Actually, I have two molecules and if I upload on maestro interface they look apart from each other, unless if translate them and put closer manually. What I am interested is the optimization of the complex of the two molecules and not keep them in a random orientation.