A　multiple　description　scalable　coding　based　on　motion　compensated　3D　wavelet　transform　has　been　proposed,　where　multiple　description　coding　is　combined　with　motion　compensated　3D　wavelet　scalable　video　coding.　First,　adaptive　rate　allocation　for　each　description　is　performed　according　to　the　motion　features　of　encoding　sequence,　so　as　to　control　the　redundancy　of　each　description.　The　key　information,　including　motion　vectors　and　low-frequency　subbands,　are　duplicated　into　the　two　descriptions,　and　high-frequency　subbands　are　assigned　into　two　separate　descriptions.　Different　methods　are　exploited　for　video　reconstruction　according　to　the　correctly　received　information.　Experimental　results　show　that,　compared　with　the　single　description　coding　method,　the　proposed　multiple　description　coding　method　can　provide　better　transmission　robustness.