Stemming is the process of extracting root word from the given inflection word and also plays significant role in numerous application of Natural Language Processing (NLP). Tamil Language raises several challenges to NLP, since it has rich morphological patterns than other languages. The rule based approach light-stemmer is proposed in this paper, to find stem word for given inflection Tamil word. The performance of proposed approach is compared to a rule based suffix removal stemmer based on correctly and incorrectly predicted. The experimental result clearly show that the proposed approach light stemmer for Tamil language perform better than suffix removal stemmer and also more effective in Information Retrieval System (IRS).