A field study was established in southern Iowa in 1994 to study seasonal and long-term weed population dynamics on land being brought back into production after 8 yr as part of the conservation reserve program (CRP). The study was a split-plot design with four replications; two tillage regimes, two crop rotations, and three herbicide application methods were used. Even though the tillage regime did not influence individual weed population density throughout the study, the no-till (NT) regime had more weeds compared to conventional tillage (CT). However, when weeds were grouped into categories, tillage influenced broadleaf weeds in 1994 and 1996 and total weeds in 1995. Plots under the NT regime had an average of 46 broadleaf weeds m−2 compared to 27 in CT in 1994, with Amaranthus rudis Sauer (common waterhemp) being the most prevalent. NT had a total of 186 weeds m−2 compared to 125 m−2 weeds in CT in 1995; however, in 1996, CT plots had 184 weeds m−2 compared to 121 m−2 in the NT regime. Except for broadleaf weeds in 1994, crop rotation did not influence the number of weeds, and herbicide application methods had the greatest effect on weed populations. Overall, weed populations were greater in 1997, 1996, and 1995 than in 1994 for all herbicide application methods. The no-herbicide treatment had the highest number of weeds throughout the study. The total number of weeds in band and broadcast treatments averaged 41 and 26 m−2 in 1994; 96 and 24 m−2 in 1995; 96 and 12 m−2 in 1996; and 109 and 95 m−2 in 1997. The use of broadcast herbicides in NT should be recommended for land coming out of CRP. Regardless of the herbicide application method or crop rotation, CT plots had better yields for both Zea mays L. (corn) and Glycine max L. (soybean). Glycine max had a better stand compared to Z. mays in the first year, indicating that a rotation starting with G. max might be preferred in the land coming out of CRP.