You can come at mRNA from two directions. Remember that human genes are largely monocistronic, in other words they often code for a single protein. If we look at prokaryotes, however, polycistronic genes such as those in operons are very much possible. From a single lac mRNA you can get multiple proteins.
For humans with monocistronic genes, you can code for multiple isoforms, however. A good example is protein kinase C: You have PKC-α, PKC-δ, and so forth. So how can this be? In part, you can look at splicing. What gets initially transcribed from your human gene is a pre-mRNA, in that the entire gene contains introns and exons. The spliceosome knows where to cut out introns, and you technically get a mature mRNA that is made of only exons and codes for protein.
In special cases, however, RNA-binding proteins and—at times in concert with, at times to the detriment of—the spliceosome machinery alternatively splice some other combination of introns and exons (sometimes, introns are kept!). And so from the PKC gene, you still get PKC, but a different isoform of it. The different proteins that different types of cells or tissues express that can interact with the splicing machinery are capable of taking the same genetic information, the same pre-mRNA, and make different polypeptides out of it.