The three different batches of an oral sustained release melatonin (MT) delivery system were prepared by aqueous-based fluid-bed coating of the sugar spheres for the evaluation of in vitro release characteristics and plasma concentration profiles in human subjects. The MT contents in 20% coated sugar spheres of three batches (B1, B2 and B3) were $3.3{pm}0.08$, $2.4{pm}0.1$ and $2.5{pm}0.13$ mg per gram of coated sugar spheres, respectively. The release profiles of three different batches had a very similar fashion. However, the release profiles of three different batches had a very similar fashion. However, the release half-lives $(T_{50%})$ of MT from B1, B2 and B3 was $3.70{pm}0.2$, $5.2{pm}0.2$ and $4.9{pm}0.07h$, respectively. Plasma concentration profiles of sustained release 0.2mg melatonin-loaded sugar spheres containing 10% immediate release melatonin in gelatin capsules (B1 and B2) were then evaluated in human subjects. The in vivo plasma concentration profies of the two batches (B1 and B2) were very similar each other and located between the physiological endogenous ranges. The time to reach the peak concentration $(T_max)$ was more advanced in case of B1 when compared to B2. However, there was no statistically significant difference in the maximum concentration $(C_max)$ and the area under the curve (AUC) between B1 and B2. The AUC of melatonin-loaded sugar spheres containing 10% and 20% immediate release MT in human subjects had a good linearity between dose and AUC, regardless of the fraction of immediate release MT, indicating the first order elimination process of MT within these doses. The current oral sustained release MT delivery system may be utilized to treat circadian rhythm disorders if it is proven to be more clinically useful when compared to immediate release MT.