Abstract: Pipeline risk assessment techniques have witnessed rapid advancement in the past few decades, but still carry on the traditional way of risk evaluation which involves risk prediction for the concerned object according to understanding on external risk based on expert's risk awareness, accident probability statistics, the core problem of the method is the risk assessment model, which is based on the concept of evaluating the whole system's risk by limited data. With the pipeline industry entering into the age of Big Data era, Big Data of pipeline has been formed under the condition of pipeline properties alignment and included in database. With the "sample equal to overall" basis in many aspects established, the risk assessment techniques and management have changed accordingly. According to engineering practice, this paper discusses the basic concept and definition of Big Data, explores and presents the possibility of replacing risk assessment with engineering applicability assessment, replacing external cause assessment with internal cause assessment, and replacing expert risk assessment with data risk assessment, and meanwhile, issues that need to be noted and developing direction are pointed out.