The Swahili had very little in terms of a proper military force, so yes, the Chinese could have decimated them if they so wished (which they rather clearly did not).
I was thinking more that China wouldn't get reinforcements the way Portuguese feitorias did -- but that point about not wanting to decimate the Swahili does play into my "why conquer" tangent.
As a materialist, that's where I get hung up -- why mutiny against Zheng He, why conquer the locals -- and why do this without the guaranteed support of the Chinese metropole? The economic incentives don't seem to be there.
I guess there's the tenuous Maqdishu idea, but that seems like a thin thread.
As for Ceylon, they'd run into the same problems that the Ports did against the interior states -- if they could conquer anything, given the differences between a Ceylonese kingdom and the Swahili cities.
Where you are getting this idea that the Portuguese "held them temporarily at best"? Mombasa was held directly for a century, and I argue the Portuguese could have held it earlier on since they thoroughly routed its Sultan in 1505 and sacked the town before leaving it destroyed. This happened twice more in the next few decades. Similarly, Kilwa was destroyed three times and the factory the Portuguese established there was abandoned only when it was a waste of money (being in an inconvenient way to India); the Portuguese left almost for good only when forced out by other foreigners (Oman) in the 1690s, and would briefly return to Kilwa in the 1720s.
There were a number of revolts, outside attacks, and temporary repulsions -- kinda like Europe's initial attempts to occupy Barbary ports outside of Ceuta and Melilla. Particularly the rising of 1631. And the Omani raid/sack of 1661.
As for Kilwa, I had forgotten about that city's destruction.
Where could they conquer that would make them to be not dependent on the sea, at least in its initial phases?
I really can't say. The easiest targets -- Taiwan, Philippines -- are all unconnected by land to the Asian mainland. Going down to Indonesia isn't much better, and would probably be just as hard if Srivijaya, Majapahit or the Chola (distantly) are still in their apogee.
Indochina doesn't offer much either -- Vietnam is too close, and at the time the Khmer were still fairly powerful, along with the Champa.