Enhancement of document images is a very essential and important stage in document image analysis, otherwise, recognition accuracy will be reduced. Hence in this work, we have proposed an effective approach to enhance the document images i.e. using fuzzy logic and a machine learning approach. The performance of the proposed method is compared with the other enhancement approaches i.e. spatial domain and frequency domain, here we have considered seven quantitative methods to evaluate the performance. Experimentation carried out on five different datasets each containing 1000 document images. Results show that the proposed method achieves good results.