This paper investigates the issues of the dynamic power management of electrical system in power assisted electric bicycle. The features of this bicycle focused on different road conditions and terrain slope when rider is cycling under the given command velocity. The proposed controller estimated the corresponding human power that the rider required and predicted the limitation of maximum assisted power from battery. With the collocation of thin in-wheel brushless dc motor and Lithium iron phosphate (LiFePO4) battery, we approached two assisted power ratio during the cycling, and observe the actual batteries power output. The performances of dynamic power management controller are experimentally evaluated to demonstrate the effectiveness of power-assisted bicycle in various cycling circumstance.