The invention discloses a task scheduling method for big data. The method comprises the following steps of: creating a task template according to a task type; Selecting a task template and a task scheduling service number to create a task, forming a configuration file, enabling the created task to comprise a task name, task content and a task execution period, configuring the task content in a kvvalue mode, and establishing a dependency relationship among the tasks through a kv file format; The task is read, a task example is generated, and task execution process information is acquired. Theinvention further discloses a big data task scheduling system. According to the big data task scheduling management method, big data scene scheduling, distribution and execution and unique characteristics of different data types are combined, the characteristics of classification, multi-triggering, strategy calling, blood relationship dependence and the like are highlighted, big data internal ecology construction is carried out, and big data task scheduling management is carried out according to different task types and different strategies through big data task scheduling.