The invention discloses power grid multidimensional sensing and a safety assessment system and method on the basis of big data. The system comprises: a power grid big data storage platform, a power grid big data security feature extraction platform, a power grid safe operation knowledge base analysis platform and a result display platform. The system adopts Hadoop file distributed management system and performs quick storage and query to big data collected by EMS; and then, power grid main security features, ancillary features and data features are extracted by the feature extraction platform and dimensionality reduction is performed to the power grid big data; the extracted security features enter the power grid safe operation knowledge base and are compared with relative safe rules in the base to obtain a safety status assessment of the current power grid. Compared with common methods, the provided system and method has following beneficial effects: an assessment system suitable for big data is designed, reasonable dimensionality reduction of power grid big data is realized, and a comprehensive and reasonable assessment for the power grid safety status is realized.