Tropical cyclones (TCs) are weather systems with vast destructive power. To give early TC warnings, accurate location of their circulation centers, or “eyes”, is required. The pattern matching solution to this TC eye fix problem works by analyzing individual remote sensing images as if they are independent. Temporal information is ignored in such approach and results from individual image are only smoothed afterwards. In this paper, a TC eye fix method using genetic algorithm (GA) with temporal information is proposed. This method gives an improvement of 35% in speed and 20% in accuracy compared to the one without the use of temporal information. Our proposed method can process 6 minutes of radar data in about 10 seconds on a notebook computer, meeting practical real time constraints. It gives an average accuracy within 0.107 to 0.156 degrees in latitude/longitude on the Mercator projected map, well within the relative error of about 0.3 degrees given by different TC warning centers.