Not until you explain what you have written to us! You can create "equivalence classes" from any set of objects- and for almost all sets, "addition" is not even defined. Are you talking about equivalence classes of pairs of numbers? Integers, rational numbers, or real numbers? And what do you mean by the sum of two equivalence classes? That has to be defined separately. Even then you can create equivalence classes with different properties by using different equivalence relations.
I
think you mean the equivalence class, defined on the set of positive integers by "(a, b)~ (c, d) if and only if ad= bc" but then the sum of two equivalence classes is
defined by the formula you give, you don't "prove" it. What is useful to prove in this case is that the "addition is well defined". That is, if (a, b) and (x, y) are in the same equivalence class and (m, n) and (p, q) are in the same equivalence class then (an+bm,bn) and (xq+ yp, nq) are in the same equivalence class.
[(a,b)]+[(m,n)]=[(a+m,b+n)]
You can define "sum of equivalence classes" to be whatever you want as long as it is "well defined". This sum might be "well defined" for a different equivalence relation. What equivalence relation are you working with?
I think you need to go back and review the basics of "equivalence relations" and "equivalence classes".