The invention provides an effective method for solving panoramic power grid multivariate heterogeneous data fusion based on big data. The big-data-based panoramic power grid multivariate heterogeneousdata fusion method comprises the steps of: defining communication protocols of a conventional wide area measurement system, a fault recorder and a light wide area measurement system, and clarifying formats and types of respective measurement data; performing shunting decoding on data frames monitored by a master station, extracting, storing, cleaning and standardizing the data in a memory, and then managing large-scale heterogeneous data in a multi-machine distributed non-relational database; and finally, performing online analysis and offline knowledge discovery of big data on a bottom layerdata interface. The big-data-based panoramic power grid multivariate heterogeneous data fusion method has the beneficial effects of realizing the effect of comprehensively analyzing three power gridmonitoring systems in a unified manner under a platform frame through large-scale heterogeneous data fusion, realizing panoramic dynamic behavior monitoring of power grid multi-voltage levels, fully utilizing the existing equipment and resources to reduce overlapping investment, and improving the operation and control level of the power grid.