Term frequency-inverse document frequency (TF-IDF) is one of the repeatedly used term weighting methods, which assigns weights based on the occurrences of a term in a document. This paper proposes an improved TF-IDF method using multi term occurrences in a document. To achieve the best performance, pre-processing methods such as tokenization, stopword removal and stemming are applied on both user query and document terms. The experimental results of the proposed work are compared with existing term weighting methods such as TF, IDF, TF-IDF and entropy. The proposed method gives better average precision, recall and F-score values than the existing methods.