x86 assembly programming loops with ecx and loop instruction versus jmp + j<condition>

i'm currently learning x86 assembly language and wondered what is the better way for implementing loops. One way would be to mov a value to ecx register and use the loop instruction and the other way would be using a jmp instruction and then comes the loop body and then a conditional jumping eventually to the beginning of the loop body. I guess the first one will has a better readability but other then that i don't know why to use it.