Research has been conducted to identify cystic mass and non-cystic mass in ultrasound images. A tota of 127 images measuring 21x21 pixels, 82 images measuring 35x35 pixels, and 78 images measuring 21x21 pixels, 82 images measuring 35x35 pixels, and 78 images measuring 50x50 pixels are taken as samples. each image was transformed into a grey-level run-length matrix and a grey-level co-occurrence matrix. there were 11 features extraced from grey-level run-length matrix and eight features extraced from grey-level co-occurrance matrix, so that totally we have 19 features. the ability of features in distinguishing cystic mass and non-cystic mass images was determind by discriminant analysis, using statistical software package SPSS version 11.5. As a result, the 19 features extracted from grey-level- run-legth matrix and grey-level co occurrance matrix could distingushing cystic masses from non-cystic mass with an accuracy of 87.3%(for image size 21x21 pixels), 91.5%(for image size 35x35 pixels), and 94.9% (for image size 50x50 pixels). further analysis carried out by involving only 12 of the 19 features extracted, which consists of 5 features extracted from GLCM matrix and 7 features extracted GLRL matrix. the 12 selected features are: energ, inertia, entropy maxprob, inverse, SRE, LRE, GLN, RLN, LGRE, HGRE, and SRLGE. Discriminant analysis with the 12 features as predictors can distinguish cystic mass image and non cystic mass with a level of accuracy of 85.3%(for image size 21x21 pixels), 91.5%(for image size 35x35 pixels), and 92.3%(for image size 50x50 pixels). further analysisi showing that area under the receiver operating curve was 0.863 (for image size 21x21 pixels), 91.5% (for image size 35x35 pixels), and 0.995(for image size 50x50 pixels), which means that the accuarcy level of discrimination is good or very good. based on that data, it conluced that texture analysis based on GLCM and GLRLM could distinguish cystic mass image and no-cystic mass image with considerably good results.