Software estimation is a skill and like all skills the more you use it the better you get at it. I would recommend reading Software Estimation: Demystifying the Black Art as it provides several different ways to approach estimating. Also I would start estimating every task and keeping track of every actual time spent. This will provide you will valuable practical experience. When I started estimating software for contract work I broke down the application into smaller pieces and created estimates for those pieces, tallied up all that time and doubled it. As the project went on I realized some tasks took less than double and some took more than double. Over the long haul I got better at it.

You might also want to reach out to the creator of Limit Theory. This was a Kickstarter that I supported in Dec 2012, which had an expected delivery date of Jan 2014. He's still working on it and has been providing regular updates. He estimated a year and it's going on 2. Remember that games that a long time.