Microemulsions (ME) are stable nano-sized emulsions that constitute one of the most effective ways to incorporate lipophilic active compounds into water-based food matrices. Great stability, colloidal domain droplets, and optical isotropy are some of the advantages of ME. The present work analyzed the influence of rheological behavior and microstructural characteristics of food microemulsions on the release rate of lipophilic vitamins E and D. A fluid ME was compared to a ME with a higher dispersed phase content and a bicontinuous structure (gel-ME). Additionally, carboxymethyl cellulose (CMC) was added to fluid ME (3.5 and 10 g/100 g), obtaining systems with different apparent viscosities and microstructure. The percentage of released vitamins was determined by HPLC, and mathematically modeled. All ME were characterized using TEM microscopy FTIR, rheological and viscoelastic analysis. Fluid and gel-ME reached higher percentages of vitamin release in a very fast manner, while systems containing CMC showed a matrix-driven nature of the release. Those systems with similar zero-shear viscosity and different microstructure exhibited significantly different viscoelastic behavior. Microemulsions with CMC exhibited a viscoelastic solid type behavior with G’ > G’’. Mechanical spectra were satisfactorily fitted with the Generalized Maxwell model and relaxation time spectra were determined. In the quiescent state, gel-ME exhibited a higher plateau modulus than those containing large amount of thickener in the continuous phase. The relaxation time behavior of the structure could explain the kinetic release of liposoluble vitamins from the inner lipid phase of the microemulsions. Display omitted •Fluid and gel microemulsions (ME) were compared with ME including a thickener.•Rheology, TEM microscopy, and FTIR were used to fully describe the analyzed systems.•Mathematical modeling of vitamins E and D release kinetics, determined by HPLC.•Vitamins release rate depends not only on ME rheology but mainly on microstructure.•Viscoelastic relaxation time spectra modeling explains differences in release rate.