No, it isn't.
The trick exploits using an interrupt call to write specific addresses of interrupt return to stack and then treat the stored addresses as commands. That is very specific to vonNeuman architecture where data, code and stack are all the same memory and what a moment ago was an address, can be reused as a command.

In 8052 stack is inside the scratchpad RAM which is completely separate address space than code memory; no amount of writing to the stack will ever change the contents of the code memory. And even if it was, '51 has no suitable interrupts; there are two interrupt levels (disabled by default anyway) and no new interrupt can be created until the previous one of the same or higher priority didn't return (removing the stack contents). What could be counted as NMI is reset which doesn't store anything on stack...