The invention relates to a kit for predicting lung cancer metastasis, and a usage method of the kit. The kit comprises a DNA library kit which comprises probes of multiple genes, and the multiple genes comprise: high-risk genes: BRCA1, BRCA2, PALB2, ATM, ATR, MUTYH, EMSY, ERCC4, RAD51, PARP1 and XRCC1; and low-risk genes: ATRX, BRIP1, FANCA, FANCB, FANCC, FANCD2, FANCE, FANCF, FANCG, FANCI, FANCL, FANCM, FANCO, FANCP, MDM2, MDM4, MLH1, NPM1, PP2R1A, PRKDC, RAD50, STAG2, XRCC5, XRCC6. Through performing related mutation detection to the related genes of DNA repair system signal paths, lung cancer metastasis can be rapidly and conveniently judged and predicted by combining a special rating mechanism.