The　precision　of　the　contact　model　for　a　joint　interface　strongly　depends　on　the　fractal　dimension　and　fractal　roughness　coefficient.　In　this　paper,　an　improved　deep　neural　network　method　was　adopted　to　predict　the　surface　appearance　parameters.　In　order　to　meet　the　high　accuracy　requirements　for　the　prediction　results　of　the　contact　model,　a　novel　surface　appearance　prediction　model　was　established　utilizing　a　regularized　deep　belief　network.　The　Bayesian　regularization　strategy　was　used　to　reduce　the　network　weights　during　unsupervised　training,　which　can　effectively　restrain　the　contribution　of　unimportant　neurons.　This　allows　to　limit　the　occurrence　of　overfitting,　and　the　layer-by-layer　training　was　performed　for　each　hidden　layer　based　on　a　continuous　transfer　function.　Meanwhile,　the　surface　appearance　parameters　of　the　joint　interface　could　be　obtained　by　plugging　arbitrary　machining　parameters　into　the　training　model.　The　specific　contact　model　was　then　established　based　on　fractal　theory　by　applying　the　above-mentioned　prediction　results.　The　parameters　of　the　joint　interface　were　used　to　simulate　the　frequencies　and　vibration　modes　of　frame-shaped　structural　parts.　The　contact　model　was　validated　by　comparing　the　simulation　results　with　experimental　data.　The　proposed　model　is　expected　to　provide　a　theoretical　basis　for　optimizing　the　structure　and　improving　the　accuracy　of　computerized　numerical　control　machines.