@ecterrab 

IE 11 may be suitable for browsing this webpage, because those picture you’ve posted can be seen by IE 11, instead of Chrome.

@ecterrab 

Thanks so much for your so detailed reply and very useful advice.

I should have ever replaced the U_0 in the ODE with any specific number to make this bug easier

to see, just as below:  

@Markiyan Hirnyk 

Please note that when U0 is nonzero(U0 isn’t U(0)), U(t)=0 isn't the answer to the ODE.

So can this be a hidden bug of Maple?

@Markiyan Hirnyk 

Thanks for your reply!

But according to the Help documents, Maple should be able to handle this problem with "dsolve({ics, ode})" in a single step.

Besides, there is still another more graceful way to tackle it:

U(t) = U__0*(exp(2*t)-2*exp(t)+1)*exp(-2*t)