Crude　oil　is　an　important　strategic　resource,　and　it　is　necessary　to　predict　its　price　change　fluctuation　accurately.　To　address　this　problem,　the　research　combines　QNN　with　DL　model　to　construct　a　hybrid　model.　First,　we　combine　the　principle　of　quantum　gate,　quantum　circuit,　quantum　superposition,　and　quantum　entanglement　to　design　the　structure　of　the　encoding　layer,　variable　layer,　and　measurement　layer.　Second,　we　embed　the　designed　QNN　into　DL　models　(RNN,　LSTM,　and　GRU),　optimize　hidden　layer　parameters,　and　construct　a　QDL　crude　oil　price　prediction　model.　Next,　this　research　uses　the　Shanghai　crude　oil　futures　price　data　and　its　influencing　factors　to　verify　the　prediction　effect　of　QDL　model.　The　research　result　indicates　that　QDL　model　is　better　than　the　traditional　DL　model.　The　integrated　QNN　model　can　improve　the　accuracy　of　crude　oil　price　prediction.　Finally,　we　take　QGRU　as　an　example　to　change　the　QNN　structure.　We　find　that　the　R2　of　QGRU　model,　which　consists　of　the　revolving　gate　Y,　revolving　gate　Z,　CNOT　gate,　and　Pauli　Z　gate,　can　reach　0.9492.　We　verify　the　stability　and　feasibility　of　QDL　model　by　changing　the　quantum　parameters.　Compared　with　the　traditional　GRU　model,　the　best　result　of　R2　is　improved　by　9.43　%.　We　further　validated　the　predictive　performance　of　the　model　under　different　dependent　variables　and　influencing　factors.　Therefore,　a　reasonable　QNN　structure　combination　can　improve　the　prediction　accuracy,　further　providing　a　new　idea　for　QDL　models　to　optimize　the　crude　oil　futures　market　and　risk　management.