Field trials were conducted during 2011 and 2012 cropping seasons at the Teaching and Research Farm of the University of Agriculture Makurdi, Benue State located in Southern Guinea Savanna of Nigeria. The objective of the study was to evaluate the response of some improved sweet potato varieties planted at three densities to intercropping with soybean. The experiment was a 2x3x3 split-split plot laid out in a randomized complete block design with three replications. The main plot consisted of two cropping systems [sole cropping (sweet potato, soybean) and intercropping (sweet potato + soybean). The sub-plot consisted of three sweet potato varieties (CIP440037, NRSP/05/007C and CIP440141). The sub-sub-plot treatment comprised of three sweet potato planting densities (25,000 plants/ha, 33,000 plants/ha and 50,000 plants/ha). Intercropping severely depressed the yields of both sweet potato varieties and the soybean component, such that intercrop yields were rarely above 50% of sole crop yields, irrespective of the sweet potato variety used. The number and weight of the sweet potato component was not significantly affected at the planting density of 50,000 plants/ha. Indices used to measure intercrop advantage showed that intercropping these sweet potato varieties with soybean was biologically efficient and percentage land saved varied from 23.08 and 32.43. Soybean was more competitive than sweet potato at all densities tested.