You must conduct an energy minimization step before moving towards molecular docking. As you have completed a homology modeling of your protein, its structure is not in a correct form currently. So, if you want to use that protein for further study e.g. Docking, MD etc, you must minimize the structure and take a stable conformer.
Energy minimization can be a good diagnostic tool - carefully compare your initial model to the minimized one. Unusually large differences show where your original model is flawed. Try to identify and correct the problem spots in the raw model, then minimize again. Do not rely on minimization alone to fix flaws in your model, in most cases, minimization alone will not improve your model, it will just hide the flaws!