Depends on your RT kit, really, though some general restriction apply.
mRNAs are long and almost always poly-adenylated at the 3' end, so oligo dT primers will bind to mRNA. Random hexamers (or similar) will also readily bind, albeit randomly (and will also bind to non-polyA RNA, like ribosomal RNA).
What you're left with is the reverse complement of your mRNA, but in DNA form. Nice and simple. You can also use gene-specific RT primers if you're only interested in specific mRNAs, too.
miRs, in contrast, are short (22nt or so) and have no poly-adenylation, so are obviously more difficult to reverse transcribe, and indeed to subsequently PCR (since they are literally the size of primers).
Various methods have been developed to circumvent this. The miScript system (Qiagen) which I use* employs poly-adenylation: adding a poly-A tail to miRs then using special reverse-transcription primers incorporating some sort of commercial sequence plus oligo-dT, so instead of miRs your cDNA is comprised of "miRs plus polyA plus commercial sequence".
These are longer, so can be PCR'd more successfully and the PCR can use a universal reverse primer that binds the commercial sequence (which they sell you) plus a forward primer that is miR specific.
Simple answer really is just "mRNA is easy to make into cDNA, whereas miRNA is really difficult".
*I am not an employee of qiagen, I just use them coz they're fairly cheap and miR-work is surprisingly expensive