A huge number of Web Services accessible on the Internet has made service discovery a challenging task. Efficient way to minimise the search space is by organising Web Services into similar clusters. This paper presents an efficient Web Service Discovery approach based on Density Based (DB) clustering algorithm. Word frequency-based similarity calculation method in word score calculation is proposed. Clusters are formed using k-means and redefined using DB algorithm. The results obtained from DB is high accurate clusters of WSDL. The performance of the proposed DB approach is evaluated by comparing the accuracy of the service discovery with the existing results and found that DB outperforms than the others.