The　optimal　control　method　for　an　intersection　traffic　signal　system　was　realized　by　applying　the　method　of　optimal　control　based　on　the　Q-learning　for　switching　hybrid　systems.　A　switching　control　method　for　hybrid　systems,　based　on　Q-learning　of　reinforcement　learning,　was　introduced,　and　the　approach　was　used　to　deal　with　the　optimal　control　for　traffic　signal　control　of　an　intersection.　Firstly,　the　hybrid　model　of　an　intersection　was　set　up.　Based　on　the　action　set　of　traffic　signal　control,　the　optimal　control　strategy　can　be　obtained　by　applying　Q-learning.　Meanwhile,　the　global　optimal　control　for　an　intersection　was　realized.　The　simulation　results　in　Paramics　simulation　software　proved　the　validity　and　feasibility　of　the　method,　which　realized　the　optimal　control　of　hybrid　switching　systems.