It depends on the detector. If you are using an FID (Flame Ionization Detector), the water won't be detected, and you will get relative amounts of the alcohols to each other, and the water will not enter into calculations. If you are using a detector that can "see" water, such as TCD (Thermal Conductivity Detector), the water will be taken into account.
You can do two things:
1) Make standardized solutions of the alcohols, and use these to calibrate the response of the GC detector. When you analyze your real sample, you can use the calibration factors to calculate the real amounts, rather than against the water peak.
2) Add a known amount of an inert "reference" material to the analytical sample whose detector response has already been measured. The unknown amounts of alcohol can be compared to the reference and have their concentrations calculated by adjusting to that response factor. This is especially accurate if the reference material is of similar chemical structure as the materials being analyzed.