Upland cotton (Gossypium hirsutum L.) production in the Southern High Plains is usually limited by water and N. The most prevalent means of irrigation in this region is center-pivot, and injection and/or ground application of liquid urea ammonium nitrate is the most common N management. The declining Ogallala aquifer has led to interest in variable-rate or variable-speed irrigation using center pivots. Variability in yields, soil properties, and elevation within the typical 48-ha center-pivot fields suggests that N fertilizer use efficiency might be improved by variable-rate N applications. We conducted a 3-yr experiment on a 14-ha area within a 48-ha center pivot in a terminated-rye (Secale cereale L.) conservation tillage cotton system. The objectives were to assess lint yield response to irrigation level for different landscape positions and to compare the effects of variable-rate N, blanket-rate N, and zero N on lint yields at varying irrigation levels and landscape positions. Lint yield response to irrigation was linear in 2002 and 2003. Increased rainfall in 2004 limited irrigation response. Contrary to our hypothesis, cotton lint yield response to irrigation was not less in the bottomslope than in the sideslopes. Nitrogen fertilizer resulted in greater lint yields in all 3 yr, but the magnitude of the response was less than that of irrigation in 2002 and 2003. Nitrogen response did not interact with landscape position or with irrigation rate. Variable-rate N resulted in more consistent lint yield response than did blanket-rate N in all years. However, when the costs of implementing variable-rate management were considered, the dollar returns to N fertilizer were favorable for variable-rate fertilization in only 1 of 3 yr.