Recent agronomic and soil science research draws attention to the importance of continuous conservation tillage (CCT), as many environmental benefits of this conservation tillage are realized only when it is used continuously over a period of years. However, little is known about the dynamics of farmers’ tillage choices. To address the need for quantitative estimates of time patterns of tillage practices and the factors that affect the use of CCT, the paper to be presented develops and estimates a dynamic model of bundled tillage-crop choices for the state of Iowa. We develop a first-order, four-state Markov chain model of tillage-crop dynamics for corn and soybean production systems. We assume that matrixes of transition from one tillage-crop state to another could vary by county but remain stationary from 1992 to 1997, and use quadratic programming to estimate the transition matrix for each of the 99 counties in the state using county-average, year–and crop–specific tillage data from Conservation Technology Information Center for 1992-1997. Analysis of Variance of the estimated county-specific transition matrixes shows that CCT occurs more often on Highly Erodible Land (HEL) when compared to other cropland. Also, the county-average probabilities of rotational conservation tillage (RCT), i.e., the farming systems in which CT is rotated with conventional tillage systems, are higher in the counties that have higher proportion of HEL. In addition, we identified a significant effect of crop rotations on tillage dynamics: the cropland under corn monoculture is less likely to be in RCT when compared to land in corn-soybean rotation. The results of the study indicate that both natural conditions (soil erodibility) and other economic choices (crop rotations) affect farmers’ choices of CCT and RCT in Iowa.