You still need to address the third issue I had mentioned. Beyond that there seems to be a difficult issue: getting reasonable starting approximations for the rate constants might be key to getting convergence in the fitting process. If the values used in FindFit are such that NDSolve cannot go to t=30 then the result you are seeing is not surprising (and the warning messages are quite to the point about this).

One possibility is to recast as an optimization problem and set NMinimize loose on it. That might give values that you can work with either directly, or as inputs to FindFit.

I placed the code in a code box. The asterisks are visible now.

See http://community.wolfram.com/groups/-/m/t/270507 , "How to post and use Wolfram Community".

There are a few issues.

(1) You do not explain what is going wrong. That really belongs in the post.

(2) There seem to be a bunch of multiplication operators missing, e.g k1s[t] should be k1*s[t].

(3) If I follow correctly you are letting t range from 0 to 30 but only solve up to t=6.

There may be other issues lurking but those were the ones I could readily find.