Objective: To test the hypothesis that the CaF2 content in a resin matrix and environmental pH significantly affect the fluoride release rate from a resin-based composite. Methods: A light-curable resin containing 70 wt% urethane dimethacrylate and 30 wt% triethyleneglycol dimethacrylate was used to make three composites containing an additional 10, 30 and 50 wt% CaF2 filler, yielding concentrations of 9, 23 and 33 wt% CaF2. Twenty discs (10-mm diameter x 2-mm thick) were made from each resin. Washed, and individually placed in 10-mL vials. Half of the vials were filled with 5 mL of pH 4.0 acetate buffer and the remaining vials were filled with pH 6.0 acetate buffer. All specimens were stored in a 37°C water bath. The buffer solution was replaced periodically up to 815 h. The ionic strength of the replaced solution was measured with a fluoride-ion-selective electrode. The cumulative fluoride release data (Y in μg/cm2) were fit to Y = a·time / (time+t½)+ b·(time)½ to estimate the values of a, the quantity of short-term ions released, t½, the time needed to release 50% of a, and b, the coefficient for long-term Fickian release. Results: The negative values of a indicate a time lag, and the values of t½ were small (<5 h), suggesting an insignificant short-term release. The mean (SD) values of b (μg/h½·cm2) for 9, 23 and 33wt% were 1.83 (0.05), 8.07 (0.08) and 8.66 (0.12) for pH=4.0 and 1.48(0.02), 7.22(0.11), and 8.59(0.12) for pH=6.0. ANOVA revealed that the mean b values are significantly influenced by the filler content (p<0.0001) and pH (p<0.0001). Further analysis showed that pH has no effect on fluoride release from the 33 wt% group (p>0.05). Conclusion: Increasing the filler content above 23 wt% CaF2 does not significantly increase the fluoride release rate. Supported by NIH/NIDCR Grant DE13412.