Abstract: In offshore basin exploration,the actual hydrocarbon source rock samples are relatively limited due to the high drilling cost.It is difficult to make the overall evaluation on hydrocarbon source rock by single well.At present,during evaluation on hydrocarbon source rocks,logging data is often used to establish the relationship with the content of organic carbon.But it is difficult to quantitatively evaluate the content of organic carbon in hydrocarbon source rock only with logging evaluation method.Therefore,it is necessary to combine many methods together.In this study,we have established logging response model for the source rock.In addition,identification and evaluation on Paleogene excellent source rocks at well points in Huizhou Sag have been conducted based on resistivity method,natural gamma ray spectrometry,cluster analysis and artificial neural network theory.The evaluation results from the logging data have been verified with the organic geochemical data and geological logging data,and the suitable evaluation methods for the study area have been obtained.