The　improved　model　based　on　slack　variables　is　designed　to　optimize　a　photonic　crystal　filter.　An　optimization　model　of　the　filter　is　established,　with　the　transmission　quantity　in　a　stop-band　as　the　objective　function　and　the　transmission　quantity　outside　a　stop-band　as　the　constraint　condition.　The　improved　model　is　obtained　by　introducing　slack　variables　in　the　infeasible　region.　With　the　treatment　of　objective　and　constraint　functions,　the　optimization　problem　and　the　improved　model　are　solved　by　sequential　linear　programming.　Then　the　convergent　solution　satisfying　the　constraints　can　be　found.　The　change　of　the　objective　function　and　design　variables　with　iteration　steps　is　presented　by　numerical　examples,　and　the　influence　of　by　the　constraints　on　the　central　frequency　of　the　stop-band　is　discussed,　indicating　that　the　improved　model　is　available　and　effective　for　optimization　of　a　photonic　crystal　filter.