When doing symbolic operations, it's going to always try to give exact representations of numbers. Ratios of integers are exact, whereas floating point representations are not necessarily exact. You can convert to a more familiar numeric form if you want.
Consider:
syms v
eqn = ((40-20)/(25-20)) == ((0.0010057-0.0009996)/(v-0.0009996))
S = solve(eqn,v) % expressed as integer ratios
format long % just so we can see all the digits
double(S) % decimal representation
The result should satisfy the equation
((40-20)/(25-20)) % LHS
((0.0010057-0.0009996)/(S-0.0009996)) % RHS is exactly equal
Though note that the floating point representation isn't exactly the same.
((0.0010057-0.0009996)/(double(S)-0.0009996))
