This study advances the synthesis of crosslinked dextran microspheres (CDMs) by addressing a critical gap in understanding

real-time droplet evolution during inverse suspension crosslinking. Using optical and SEM microscopy, we characterize

droplet progression through four distinct stages—transition, quasi-steady-state, growth, and identification—revealing the role

of viscosity-dependent breakage-coalescence dynamics in size distribution changes. We further develop the first population

balance model (PBM) integrated with genetic algorithm (GA) optimization to predict transient particle behavior. By incorporating

a reaction-conversion parameter, X(t), our model links rheology to crosslinking kinetics, achieving high predictive

accuracy (MSE < 5%). Experimental results demonstrate that increasing dextran concentration (12.5–50% w/v) elevates

viscosity by > 3000%, suppressing droplet breakage and producing larger particles (27 → 188 μm) with broader distributions

(Span 0.98 → 2.61). This work represents a significant improvement over previous statistical approaches, offering the first

quantitative PBM-GA framework for connecting processing conditions to dynamic particle evolution. Our findings provide

new insights into CDM formation kinetics and enable rational microsphere design for biomedical applications, bridging the

gap between empirical observation and mechanistic control in dextran-based particle synthesis.