Disclosed is an extraction method for time-space-spectrum four-dimensional remote sensing data. The method includes: obtaining remote sensing images at a preset coverage area during a preset time period (S1); re-projecting the remote sensing images in a way that the respective pixel positions of the respective remote sensing images at different time points are overlapped (S2); storing all of re-projected remote sensing images into a storage unit according to a first preset storage method, or storing all of the re-projected remote sensing images into the storage unit according to a second preset storage method (S3); determining, on the basis of requirement information for data extraction, whether current data format of remote sensing data is a data format matching the requirement information (S4); if yes, determining, on the basis of the requirement information, a storage location of required remote sensing data, and extracting the required remote sensing data (S5). In the case storing is executed through the two preset storage methods above, the time consumption for locating the storage location can be reduced, thus the efficiency in processing the four-dimensional remote sensing data can be improved.