The quality of tropical cyclone (TC) best track data has increasingly received more attention in climate research in recent years because of the rising interest in the possible effects of global warming on TC activity. In this paper, three best track datasets for the western North Pacific TCs were compared in a seasonal context. Firstly, the statistical characteristics were examined based on their distributions. Then, the potential impact of particular features within the datasets was evaluated by using contrastive cluster analyses. The mean amplitude of the annual cycle in the Japan Meteorological Agency (JMA) data is largest among the three datasets. The smallest interannual variability helped the Shanghai Typhoon Institute of China Meteorological Administration (CMA-STI) dataset to resist the negative effect of outliers. An inhomogeneity around 2003–2004 was identified in the Joint Typhoon Warning Center (JTWC) data by the outliers and was compared with the JMA and CMA-STI data. By using contrastive cluster analyses, the differences among the datasets and some effects related to particular issues, including the representation of the annual cycle, the outliers, and wind conversion were evaluated. The JTWC dataset is affected by outliers and wind conversion among the different average times, whereas the representation of the annual cycle was important for the CMA-STI data. After excluding all the outliers, the results indicated more consistent classifications of the TC annual cycle between the JMA and JTWC data.