Continuous intensive monocultures of rice can lead to increased soil compaction, reduced soil quality and decline of rice yield. The objectives of this study were to evaluate the effect of rotations of rice with upland crops on soil quality of an alluvial clay, and on yield and yield components of rice in a long-term experiment in the Mekong Delta. A field experiment was established in 2002 for 10 years at Cai Lay district - Tien Giang province using a randomized complete block design with four treatments and four replications: (1) rice-rice-rice (control - conventional system as farmers- practice), (2) rice-maize-rice, (3) rice-mung bean-rice, and (4) rice-mung bean-maize. Prior to our experiments, the experimental site was cultivated for over 30 years under a conventional system. The soil samples were collected at three depths 0-10, 10-20 and 20-30 cm. Results showed that the rotation of rice crop alternated with upland crops contributed to a significant increase in soil organic matter, soil labile carbon, soil available nitrogen and improved soil bulk density, soil porosity, soil aggregate stability index and soil strength in comparison with traditional rice mono cropping practice. As a consequence, rice plant height, total tillers, panicle, filled grain percentage and yield was improved when rotating rice with upland crops in comparison with the conventional rice monoculture practice: rice yield increased compared with the control with 32 to 36%. Our study demonstrated that rotating rice with upland crops is recommended to improve soil quality, rice productivity and farmer-s income hence contribute immensely to sustainable agriculture. Therefore, we recommend large scale implementation of these rotations on alluvial clay soils and possibly on other soils in Mekong Delta.