The embodiment of the invention provides a user behavior analysis method and system based on streaming calculation. The method and system are used for achieving real-time analysis of user behaviors and improving the accuracy of user behavior analysis. The method comprises the steps that user behavior data is acquired from at least two channels and is converted into JSON data packets in a unified format, wherein the user behavior data comprises user behavior paths, time indexes, user basic label data, transaction data, activation and retention, click behaviors and platform types; a Spark Streaming computing engine is used for carrying out real-time user behavior analysis on the user behavior data, wherein the user behavior analysis comprises event analysis, user retention analysis, funnel analysis, user behavior path analysis, transaction analysis, user portraits and click analysis; user behavior analysis results are stored, and corresponding analysis results are displayed according toquery requests of users.