In this paper, a classification method for in-shell hazelnuts and shelled hazelnuts is proposed using image processing algorithm based on thresholds. Hazelnuts are sold as in-shell type with an outer fibrous husk, shelled type with dark brown skin, blanched type without skin and roasted type which means roasted seed after peeling for skin. However, outer husk is easily peeled during production, convey and processing. Therefore, classification for in-shell hazelnuts and shelled hazelnuts is required to sell hazelnuts according to processing steps. In this paper, a classification method for in-shell hazelnuts and shelled hazelnuts is proposed using 18 classification parameters based on image processing algorithm with thresholds. As a result of classification for hazelnuts, discrimination accuracy for in-sell hazelnuts was 98% and discrimination accuracy for the remain shelled hazelnuts was 94%.