Provided are a method and apparatus for encoding and decoding in an inter mode based on a multiple scanning through which efficiency of the inter mode is improved by using a spatial relationship within a frame and obtaining estimation information from numerous references. The method includes dividing an input image into basic encoding/decoding units including predetermined sized blocks, scanning at least one of the basic encoding/decoding units, determining whether references for motion estimation of a current basic encoding/decoding unit can be used as references to encode/decode the current basic encoding/decoding unit, and selectively encoding/decoding the current basic encoding/decoding unit according to a result of the determination. A decoded previous frame or blocks decoded before the final scan of a current frame is used the references for motion estimation.