Hydrocarbon prediction is a crucial issue in the oil and gas industry. Currently, the prediction of pore fluid and lithology are based on amplitude interpretation which has the potential to produce pitfalls in certain conditions of reservoir. Motivated by this fact, this work is directed to find out other attributes that can be used to reduce the pitfalls in the amplitude interpretation. Some seismic attributes were examined and studies showed that the attenuation attribute is a better attribute for hydrocarbon prediction. Theoretically, the attenuation mechanism of wave propagation is associated with the movement of fluid in the pore; hence the existence of hydrocarbon in the pore will be represented by attenuation attribute directly. In this paper we evaluated the feasibility of the quality factor ratio of P-wave and S-wave (Qp/Qs) as hydrocarbon indicator using well data and also we developed a new attribute based on attenuation for hydrocarbon prediction -- Normalized Energy Reduction Stack (NERS). To achieve these goals, this work was divided into 3 main parts; estimating the Qp/Qs on well log data, testing the new attribute in the synthetic data and applying the new attribute on real data in Malay Basin data. The result show that the Qp/Qs is better than Poisson's ratio and Lamda over Mu as hydrocarbon indicator. The curve, trend analysis and contrast of Qp/Qs is more powerful at distinguishing pore fluid than Poisson ratio and Lamda over Mu. The NERS attribute was successful in distinguishing the hydrocarbon from brine on synthetic data. Applying this attribute on real data on Malay basin, the NERS attribute is qualitatively conformable with the structure and location where the gas is predicted. The quantitative interpretation of this attribute for hydrocarbon prediction needs to be investigated further.