The　scattering　of　fatigue　life　data　is　a　common　problem　and　usually　described　using　the　normal　distribution　or　Weibull　distribution.　For　solder　joints　under　drop　impact,　due　to　the　complicated　stress　distribution,　the　relationship　between　the　stress　and　the　drop　life　is　so　far　unknown.　Furthermore,　it　is　important　to　establish　a　function　describing　the　change　in　standard　deviation　for　solder　joints　under　different　drop　impact　levels.　Therefore,　in　this　study,　a　novel　conditional　probability　density　distribution　surface　(CPDDS)　was　established　for　the　analysis　of　the　drop　life　of　solder　joints.　The　relationship　between　the　drop　impact　acceleration　and　the　drop　life　is　proposed,　which　comprehensively　considers　the　stress　distribution.　A　novel　exponential　model　was　adopted　for　describing　the　change　of　the　standard　deviation　with　the　impact　acceleration　(0　-＞　+a).　To　validate　the　model,　the　drop　life　of　Sn-3.0Ag-0.5Cu　solder　joints　was　analyzed.　The　probability　density　curve　of　the　logarithm　of　the　fatigue　life　distribution　can　be　easily　obtained　for　a　certain　acceleration　level　fixed　on　the　acceleration　level　axis　of　the　CPDDS.　The　P-A-N　curve　was　also　obtained　using　the　functions　mu(A)　and　sigma(A),　which　can　reflect　the　regularity　of　the　life　data　for　an　overall　reliability　P.