You are not going to be able to forecast oil and gas prices with any useful accuracy at all. Lots of people try, and as far as we know, every one of them has failed. If anyone has ever succeeded, then they've become extremely rich through it, and they're not telling anyone the secret. If you did succeed, then your energy economics model would be completely irrelevant to you, as you'd be too busy trading the oil market, and eyeing up which tropical island you should buy next.
The best you can hope for is consistency in your model. And then sensitivity-test your model based on a broad range of assumptions.
For example, my usual 95% credible range for the price of a barrel of oil in 2030 is USD(2016) 5-500.
So, what do you look for, for consistency? Demand and supply. Have short-run and long-run curves for both. If you're forecasting more than five years ahead, you must have some kind of uptake model for substitutes too, as well as short-run and long-run supply curves for them. Depending on how complex your model is, you might want to model storage too, particularly if modelling in less than 5-year time steps.