Abstract: The content of asphaltene, resin and wax in crude oil is an important basic data for oilfield development, process design of crude oil production and transformation and operational planning. However, the shortcomings such as long cycle, low precision and some solvents with toxicity and harmfulness always exist in the traditional standard testing methods of the group compositions, which could have seriously effect on the testing efficiency of the group compositions and the reasonableness of production technics and operational plans. Therefore, based on the principle of chemometrics, 300 groups of crude oil samples from Bohai area were tested, a multivariate calibration model was established by infrared spectroscopy and asphaltene, resin and wax content in crude oil. The model can predict the group content of asphaltene, resin and wax in the crude oil through detecting its infrared spectrumis and the contrastive repeated experiments. It is found that the average relative error, average absolute relative error and standard deviation of the method are 5.3%, 5.1% and 6.1% respectively, and its repeatability and reproducibility are less than 3% and 15% respectively, which is superior to those of the standard detection method. Compared with standard methods, this method has shorter testing time, lower cost and higher reliability and testing efficiency.