Cancers in head and neck tends to spread to nearby lymph nodes. Lymph nodes trap the spreading tumor cells but then the tumor starts to grow in these nodes and then spread further. Our project is aimed to (1) predict the secondary regions of lymph nodes where a given primary tumor can metastasize, and (2) to generalize a pattern for lymph nodes metastasis of head and neck cancer by using artificial neural networks (ANN). The raw data for the analysis is provided by Dr. Lincoln Gray, Acta Otolaryngologica 2000, of 130 cases of pathologically-positive oral squamous cell carcinoma (SCC) from UK. Seven primary sites for tumors are identified: 1) buccal mucosa, 2) tongue, 3) retromolar trigone, 4) floor of mouth, 5) ventral tongue, 6) oropharynx, 7) lower alveolus. Ten secondary regions of lymph nodes metastasis are observed: five regions each on the same/opposite side (ipsilateral/contralateral) as the primary tumor site. In our oral squamous cell carcinoma study using ANN, we explore data analysis approach with two ANN methods: (1) a supervised multilayer feed forward back propagation (back-prop) method, and (2) an unsupervised self-organizing map (SOM) method. This experience provides insight into implementation of ANN and directions to future investigation. The results from back-prop are comparable to that using multidimensional scaling (MDS) with respect to prediction of lymph nodes that have highest percentage of being metastasized, while SOM requires further work to identify clustering for individual primary cancer as well as next level of lymph node metastases.