In　this　paper,　a　robust　match　algorithm　for　shadow　area　extraction　in　the　process　of　visual　autonomous　navigation　for　attitudes　estimation　in　planetary　landing　is　proposed.　First,　the　stable　shadow　areas　areextracted.　Then,　an　affine　normalization　algorithm　is　applied　to　normalize　the　shadow　areas.　At　last,　the　featuresare　extracted　by　using　multiple　angles　sift　(MA-SIFT)　description　operator,　the　matching　pairs　are　got　and　the　mismatches　are　removed.　Comparative　trial　are　conductedto　compare　with　SIFT　algorithm.　It　is　shown　that　the　matching　rate　can　be　maintained　at　a　higher　level　by　the　use　of　the　proposed　algorithm,　even　there　are　large　visual　angle　change　between　image　pairs.　©,　2015,　China　Spaceflight　Society.　All　right　reserved.