Normally, at such an early stage it is not about valuation, but rather about contribution. Each partner should get percentage that reflects their contribution to the project (say, in terms of time, skill set, contacts, and/or self-financing). There are various formulas that try to do that, but they often give different results and contain subjective factors.
In my experience, in the end it always comes down to negotiations – all potential partners decide what percentage they will feel good about.
In any case, I would suggest not to get cheap on this – good partners are hard to find. Remember that there is a ~90% chance that you will fail and these percentages will become meaningless, and if you do succeed and get millions, then does it really matter if it’s one million more or less?

Scott, an advisor in Silicon Vallyer answered:

Project your next month revenues. Get comps of similar public companies vs. their previous year’s revenue (e.g. 3x). That’s a reasonable guess about what your company will be worth next year if you are on target. There is some high probability you will fail and lose everything, some probability that you will sell less than forecast, and some probability that you will earn more. The expected value of your company is the weighted sum of all those probabilities times their values. For sake of illustration let’s say that there is a 50% chance sales will be $0 and 50% chance the company will sell an average of $1M. Take the weighted average and you get an expected value of sales of $0.5M. Multiply that by the industry multiplier (e.g. 3x) and you get an expected valuation of $1.5M next year. Technically you should adjust this by the time value of money based on current interest rates. If the values of the payoffs were well known you definitely should adjust for that, in this case the errors in size of next year’s sales, the probabilities, and next year’s multiples probably swamp the interest rate corrections so we are only getting a stake in the sand estimate for a negotiation staring point. In the end you might adjust that estimate to get agreement between both parties. But at least you now know a way to find a starting point for negotiations.

Finally, check out slicingpie.com, probably it can give you a more clear way.

I have made a lot of mistakes. Hope this post can help you.