Summary A large set of field data have been collected during the last decade related to various facets of asphaltene instability problems at our production facilities. These problems include compatibility of heavy oil with hydrocarbon diluents in a Venezuelan operation, commingling of live oil and condensate in a North Sea production facility, compatibility of drilling mud-base oil, and miscible injectants with reservoir fluids in Alaskan operations. In each of the field cases, significant lab data were generated by titrating the neat dead crude oil, and oil-solvent blends with n-alkanes. The solvents used to study the asphaltene issue were toluene, condensate, and base oil. We have applied available asphaltene prediction techniques (Heithaus 1960, 1962; Wiehe and Kennedy 2000ab; Wiehe et al. 2001; Andersen 1999; Wang and Buckley 2001; Wang et al. 2003, 2004; Leontaritis 1998) to explain the field data. None of the models has been found comprehensive enough to explain flocculation at all the conditions, including the flocculation that occurs at ambient conditions in the presence of paraffinic diluents; stability enhancement that occurs upon addition of aromatic solvents; and the instability that occurs in a live fluid because of changes in composition, pressure, and temperature. To handle a complex crude oil system, these models made some simplifying assumptions that enabled them to make the problem manageable. In doing so, they lose some predictive capability. We found there are two forces that need to be accurately captured— dynamics of the alteration of solubility parameter of the hydrocarbon matrix, and change in entropy of mixing—to model the asphaltene behavior. The latter has been either empirically estimated by extrapolating the ambient titration data or neglected in many of the previous models. The basic parameters for our model can be calculated from lab data generated by titrating the dead crude oil, or oil solvent blends with n-alkanes, at different temperatures. So far, this model has been applied to various field conditions in production facilities and has been found successful in matching the field data.