Raman spectroscopy proved itself a non-destructive technique which is being used extensively for the characterization of biomedical media for disease diagnosis. It is based on the inelastic scattering of light. It provides the information about the vibrational modes of a molecule. A Raman spectrum of a biological media contains the molecular fingerprints which are the intrinsic characteristics of molecules. This information can be utilized to train various statistical models that as a result become a powerful tool for the characterization, diagnosis and discrimination in a variety of ways. In this thesis, the outcomes of three studies based on the Raman spectra of body fluids have been presented. First of all, a study was conducted to choose an optimal substrate to be used for acquisition of Raman spectra of body fluids. Aluminum was selected as the most suitable substrate for recording Raman spectra of human blood and sera samples. In first study, blood samples of healthy and breast cancer (BRC) female subjects were used to analyse the molecular variations caused by BRC. Partial least squares (PLS) regression based model was developed, which highlighted the bio-molecular variations found between both groups. Among these molecules lycopene, phosphatidylserine, qunoid ring, calcium oxalate and calcium hydroxyapatite were found positively correlated while tryptophan, proline, valine, glycogen and tyrosine were negatively correlated with BRC. In second study, human blood sera samples were analysed for the molecular variations due the presence of antigen at an early stage of dengue virus (DENV) infection. It was found that level of cytokines and DNA rises, while the level of proteins with alpha and beta conformations along with lectins lowers in the sera due to DENV infection at an early stage. In third study, human blood sera samples were analysed for the screening of hepatitis C infection. Hepatitis C infected sera were tested for the presence of antibodies by ELISA. Analysis of regression coefficients revealed that trehalose, chitin, ammonia and cytokines are positively correlated with the hepatitis C while lipids, beta structures of proteins and carbohydrate binding proteins like MBL are negatively correlated with hepatitis C.