Astudy was carried out during 1993-1995 at the Liliana Dimitrova Horticultural Research Institute in La Habana, Cuba in order to evaluate the efficiency of cassava-bean intercropping (Manihot esculenta Crantz-Phaseolus vulgaris), and to determine the most adequate topological arrangement. The research was performed on red ferrous compacted soil. 1.4-m beds, and a randomized block design with four repetitions were used. Nine treatments were tested at three different planting times and with three different spatial distributions; control monocrops were also studied. The best topological arrangement for cassava was when it was planted at the bed center (28.3 t/ha), and beans were sowed twenty days later on both sides of the cassava. Beans also showed their best yield (1.53 t/ha) under this condition. A land use equivalent (LUE) of 1.01 – 1.85 was obtained when calculating all the combinations together.