There are two SU(3) symmetries you can come across. Basically, SU(N) emerges everywhere when you have N quantum states and some physics does not distinguish these states - then all quantum superpositions of these states make a fundamental representation of SU(N) significant to that physics (but maybe insignificant to some other).
Thus in particle physics there are two SU(3) discussed often. The SU(3) of colors represents the exact, unbroken symmetry. Then there is the SU(3) of flavours, which mixes $u$, $d$, $s$ quarks in strong interactions. It is very rough and it is broken by the masses of quarks ($m_s$ is about 150 MeV, that is dozen percent of the QCD scale of about 1 GeV). After breaking it leaves the flavour SU(2) of only $u$, $d$ quarks, and that is much more accurate ($m_d-m_u$ in the order of unities of MeV), though still approximate. Nevertheless, these symmetries are used to classify hadron states, and to descibe physics on the hadron level (which is the effective theory with respect to SM). It may be credited for the close masses of hadron multiplets, but this is (almost) another way to say "hadrons have close masses because quarks inside them have close masses".
Also, $SU(3)_C$ is a gauge (local) symmetry, while $SU(3)_F$ is a global symmetry.