Competition between crops and weeds is a complex phenomenon. Comprehen
sive, process-oriented simulation models which treat competition in a
mechanistic rather than an empirical fashion, can offer insight into r
elationships among competition, crop and weed density, relative time o
f emergence, various morphological and physiological traits, and resou
rce levels. They can also be used for prediction as part of a systems
approach to weed management. This paper reviews the features of a numb
er of recent simulation models of crop-weed competition, the species f
or which they have been parameterized, and their applications. To date
, these models have been used primarily to predict crop yield losses d
ue to weed competition. Their ability to simulate weed seed production
in response to the environment has not been exploited. The next step
is to link simulation models of crop-weed competition to weed populati
on dynamics models, in order to improve our ability to predict the eff
ect of various weed management strategies over time. Advantages and dr
awbacks of a modeling approach to weed management problems are discuss
ed.