There are many linguistic morphology tools available in the market for commercial and research purposes. Morphology technique are incorporated into these tools to ensure its ability to study the internal structure of natural language words. This technique plays an important role in reducing the number of vocabularies used, at the same time retains the semantic meaning of the knowledge in NLP system. Among the algorithms implemented, majority of them only has to ability to carry out stemming process instead of a lemmatization process. Even with technology advancement, yet none of the available lemmatization algorithms able to produce 100% accurate result. Inappropriate words produced by the current algorithm might alter the overall meaning it tried to represent, which will directly affect the outcome of NLP system. This paper proposed a new method to handle lemmatization process during the morphological analysis. The method consist three layers of lemmatization process, which incorporate the implementation of a well known Stanford parser API, WordNet database and adaptive learning technique. Stanford parser API is implemented in the first layer of lemmatization process, whereas WordNet database and adaptive learning technique are implemented in the second layer and finally another lemmatization algorithm in the final layer. The lemmatised words yields from the proposed method are much more appropriate compare to the previous algorithms due to user participation in the adaptive learning technique, which will ultimately improve the semantic knowledge represented and stored in the knowledge base.