Abstract To compare the encapsulation of annatto extract by external gelation (EG) and internal gelation (IG) and to maximize process yield (% Y), two central composite designs were proposed. Calcium chloride (CaCl 2 ) concentration (0.3–3.5%), alginate to gelling solution ratio (1:2–1:6); acetic acid (CH 3 COOH) concentration (0.2–5.0%) and alginate to gelling solution ratio (1:2–1:6) were taken as independent variables for EG and IG respectively. Release studies were conducted under different conditions; morphology, particle size, the encapsulation efficiency (EE), and release mechanism were evaluated under optimized conditions. The optimized EG conditions were 0.3% CaCl 2 and 1:1.2 alginate to gelling solution ratio, whereas a 0.3% CH 3 COOH and 1:5 alginate to gelling solution ratio were optimized conditions for IG. When 20% extract was employed, the highest EE was achieved, and the largest release was obtained at a pH 6.5 buffer. The Peppas–Sahlin model presented the best fit to experimental data. Polyphenol release was driven by diffusion, whereas bixin showed anomalous release. These results are promising for application as modulated release agents in food matrices.