Vehicular　ad　hoc　networks　(VANETs)　as　the　basic　infrastructure　can　facilitate　applications　and　services　of　connected　vehicles　and　intelligent　transportation　systems.　To　reduce　the　impact　of　deficiency　of　spectrum　resource　in　VANETs,　cognitive　radio　(CR)　technology　can　be　applied　to　VANETs.　Although　CR-VANETs　can　utilize　the　unused　licensed　spectrum　effectively,　the　distributed　nature　of　CR-VANETs　may　open　a　door　for　different　malicious　attacks,　such　as　spectrum　sensing　data　falsification　attacks.　In　this　paper,　with　the　recent　advances　in　fog　computing　and　network　function　virtualization　(NFV),　we　propose　a　hybrid　fog　computing　model　to　enhance　security　in　CR-VANETs.　Based　on　this　fog　computing　model,　we　propose　a　new　service　named　Spectrum　Sensing　as　a　Service　(SSaaS),　which　can　perform　cooperative　spectrum　sensing　in　CR-VANETs　with　fog　computing　to　secure　the　spectrum　sensing　procedure.　Simulation　results　are　presented　to　show　that　fog　computing　in　CR-VANETs　can　effectively　reduce　latency　and　improve　the　security　of　CR-VANETs.