"We (project managment team) supposed to run all three projects simultaneously"
This is probably the absolute best way to ensure that none of the projects are completed on time, or on budget. Even if your team(s) manage to waste absolutely no time context switching between projects (which is basically impossible) why deliver a project in 6 months that you could deliver in 2?
" Ofcourse all clients want their project to be done first... "
That's nice, but irrelevant to the question at hand. Clients always want their project done first.
"Now, how can we determine which project should kick in first?"
Firstly, it isn't which project should kick in first, it is which project should we complete first. This is a vital distinction. You shouldn't begin the next project until you finish the one you are on. All you will do otherwise is waste time and resources.
Secondly, you need a metric with which you can evaluate each of the projects. Cost of Delay is what you are looking for. How much money will it cost to delay project A 1 day (or week, or month or whatever scale makes sense)? Now your decision is simple. The cost of delay calculation doesn't even have to be that accurate, it just has to be measured for each of the 3 projects in a similar fashion. If all 3 measurements are off my a factor of 2 in the same direction, its still a good comparison.