Karyotyping is essential for identifying genetic ab-normalities, assisting with prenatal screening, gaining insight into the genetics of cancer, researching evolution, and advancing research. It offers useful knowledge on chromosomal anomalies and makes contributions to many fields of biology, genetics, and medicine. A key component of karyotyping is chromosome classification, which entails classifying and arranging chromo-somes according to their size, shape, and banding patterns. While Denver categorization of chromosomes offers a defined and widely acknowledged system for naming and classifying human chromosomes into 7 groups, chromosome classification categorizes the human chromosomes from the metaphase images to 23 or 24 classes. In order to improve the effectiveness of the karyotyping system, a multilabel chromosomal classification approach is developed in this study. It makes use of the combined information from the 24 class chromosome classification infor-mation and the Denver classification information. The suggested method uses two Convolution Neural Networks, one for 24 class classification and one for Denvor classification. The proposed model is experimented with the public ChromosomeNet dataset as well as a private dataset generated at Regional Cancer Center, Thiruvananthapuram, Kerala and obtained the testing accuracy of 97 % and 70 % accuracies respectively. For the comparative study, CNN features of the 24-class classifier are used to train the traditional classifiers Decision Tree, Random Forest and SVM and proved that the proposed multilabel classifier outperforms the other models.