Parallel factor analysis (PARAFAC) and alternating residual tri-linearization (ART) algorithms are used to measure and identify petroleum pollutants. The differences between the two algorithms in oil identification are emphatically compared and analyzed. The CCl4 solutions of No. 95 gasoline, No. 0 diesel and kerosene are used as the research objects. We take petroleum mixed solutions with different concentrations as samples to measure the three-dimensional fluorescence data of each sample by F-7000 fluorescence spectrometer. When PARAFAC algorithm is applied and the component number is set to 3, the recovery rates of diesel, gasoline and kerosene are (95.60±3.60)%, (94.67±3.66)% and (95.49±4.49)%, respectively. ART algorithm does not require a preset component number, and the recovery rates of diesel, gasoline and kerosene are (96.58±2.17)%, (95.17±9.17)% and (95.90±8.90)%, respectively. The results show that the two algorithms can be used for the measurement and identification of three kinds of petroleum pollutants, and high recovery rates can be obtained. ART algorithm does not require presetting component number, so it has more advantages.