Recently, memristor has been studied as one of the most promising candidates for next generation nonvolatile-memory due to its fast write/read speed, superior scalability, compatibility with CMOS technology and multi-bit storage potential [1]. We here represent the empirical current-voltage (I-V) model to explain the Pt/γ-Fe 2 O 3 /Pt (Fig. 1 (a)) memeristive switching behaviors based on the variation of the state variables. Noticeably, physical mechanisms such as tunneling [2] and back-to-back Schottky characteristics are taken into account (Fig. 1(b)). In addition, we show that the proposed empirical I-V model can be successfully incorporated into the SPICE model using Verilog-A (Fig. 1(c)). Our model has a distinguished merit in perspective of the opimization of material, process, and memeristive device in that it is able to be identified by the resistance variation of ON state in the formula and without any fitting parameter in the OFF state. This study would be helpful in understanding of the memristive behaviors and also give insight into the design for innovative memristor-based circuit applications. Fig 1. (a) Schematic of the Pt/γ-Fe 2 O 3 /Pt memristor, (b) equivalent circuit of the memristor, and (c) comparison of empirical I-V model on the measured I-V characteristics. [1] Y. Schottky diode (Between top electrode and oxide layer) Schottky diode (Between oxide layer and filament) Memristance (filament) ON state ¢ £ V3 V2 V1 (a) V app GND 儐-Fe2O3 Nano Particles Pt top electrode Pt bottom electrode Substrate (Ti/SiO 2 /Si) -6 -4 -2 0 2 4 6 -6 -4 -2 0 2 Reset process 6th 7th 8th 9th 10th Symbol : Measurement Line : Model 1st 2nd 3rd 4th 5th Current, I [PA] Applied Voltage, V app [V] Set process (c) (b)