The best way is to grow the train in the absence of any selective pressure, at least for an overnight; you then can plate dilutions on medium w/o antibiotic, to get well separated colonies, and replica plate them on antibiotic-containing medium. You can thus see which cells have lost the plasmid (growth only on the plate w/o antibiotic).
Some plasmids are easy to cure but some are not. Basically, you streak your bacterial strain harboring the plasmid on an agar plate without antibiotic whose resistant gene is on the plasmid. Or you can grow your plasmid-containing bacterium in liquid medium without antibiotic selection. I personally prefer the later. After the liquid culture is ready, you can pellet 1 mL culture and then make a serial dilution before plating on antibiotic-free agar plate. The plate gives you the best separation of colonies is used to replicate those colonies on a antibiotic-containing plate. The antibiotic sensitive colonies are your plasmid-free strains. You probably have to do the liquid growing a few times if the plasmid is hard to be eliminated. Taking 1 mL culture from first round growth and inoculate to another batch of liquid medium, growing at the same condition. Good luck.
Just realize that many plasmids are quite stable and are only lost at a very low frequency, so you might need to screen hundreds or thousands of colonies after repeated culturing without antibiotics. It very much depends upon the plasmid.