Earthquakes are sudden natural disasters that endanger people's lives and properties all around the world. Although much work has been done in the field of prevention of such damages via timely prediction the work needs to be tailored to the area-specific dataset for optimum accuracy. With the time series seismic event data of Chiang Mai for 4 years, we focus our study to provide a prediction of future earthquakes here by using z-normalization and a polynomial constraint in DTW on CUDA-enabled accelerators. By comparing polynomial constraint to Sakoe-Chiba Band as well as Itakura band, we have verified our work to have reduced the percentage of computational time and improved accuracy. The aim of this paper is to demonstrate a prediction accuracy of 0.95.