This is where I point everyone to the
Bilali document and to Professor George Junne's article
Neither Christian nor Heathen. The TL

R summary: some of the Muslims brought to British North America and the Caribbean as slaves were literate, some of those tried to preserve their faith, and there was actually sporadic communication with imams in Africa (although this was of course infrequent and unreliable). In the Gullah/Geechee country, where most of the slaves came from the same part of Africa and where they had more freedom than other slaves to arrange their off-duty lives, Islamic practices were maintained for two or three generations, and while there are no Gullah Muslims today, there are cultural survivals that can be traced to Islam.
With that said, any propagation of Islam under American slave traditions would be nearly impossible. As has been pointed out, the second-generation slaves would lose literacy and the Arabic language, and since there was little communication between slaves from different plantations, they'd have no easy way to spread the faith to their neighbors. Any such effort would also have to contend with the active promotion of Christianity by the plantation owners.
Maybe, if emancipation could somehow occur in the late 18th or early 19th centuries, followed by a period when the Gullah were essentially left alone, Islam could survive there. A group of Muslim slaves successfully going maroon might be another option. In either of those cases, though, isolation and lack of literacy would mean that folk-Islam rather than orthodox Islam would prevail.
Failing that, another POD might be if the American slave system were more like the Brazilian one. Slavery in Brazil
was different - slaves of the same ethnic group tended to be kept together, there were more urban slaves with literacy and freedom of movement, and there was a large free black community that belonged to the same religious-social brotherhoods that the slaves did. Preservation of Islam was much easier in that environment than in the United States. Outside special cases such as Gullah country, though, I'm not sure how to arrange such a change in the structure of American slavery.