The invention relates to a detection system for singular sequence data and a detection method for the same. The method conducts singular sequence data detection based on a Hadoop platform and band weight optimization. The method comprises the steps that paired comparison is conducted to a dataset, so outlier degrees are generated; generated results of the outlier degrees are sequenced, the results are gathered based on voting theories, and a comprehensive outlier degree of all the sequence data in the whole sequence dataset can be obtained; and the outlier degree is used to optimize a voting weight of each sequence in the dataset, and iterative execution is carried out till the weight result is converged or the number of iterations reach an upper limit. The technical scheme provided by the invention solves the problem that data standards can hardly be acquired, and solves a performance problem caused by a huge data size.