Objective：To establish and compare valproic acid（VPA） or lipopolysaccharide（LPS） induced rat model of autism，so as to provide a scientific basis for the selection of animal models in autism research. Methods：Rats were injected intraperitoneally with 500 mg/kg VPA or 100 μg/kg LPS on the gestational day 12.5 or day 9.5 to construct autism model in offspring rat（n=28）. The pup’s physical development was continuously monitored，and social interaction，communication ability，stereotype and repetitive patterns of behaviors were detected in the pups. ELISA kits were used to determine the concentrations of plasma inflammatory cytokines（IL-6，IL-1β，IFN-γ and TNF-α） in the pups with or without LPS treatment. Levels of PTEN，p-Akt and p-mTOR protein expressions in the prefrontal cortex of the pups in the present or absent of VPA treatment were confirmed by Western blot. Concentrations of 5-HT in the prefrontal cortex with or without VPA challenge were detected with high performance liquid chromatography（HPLC）. Results：Both of the two autism models displayed a developmental delay and autism related behavioral abnormality. Levels of IL-6，IL-1β，IFN-γ and TNF-α were increased significantly in the rat offspring following LPS treatment（t=2.108，P=0.049；t=2.312，P=0.033；t=2.809，P=0.016；t=2.110，P=0.049）. The VPA challenge significantly decreased the level of PTEN protein expression（t=13.320，P=0.006），and increased the expression levels of p-AKT and p-mTOR in the pups’ prefrontal cortex after the VPA treatment（t=4.633，P=0.044；t=7.984，P=0.015）. The concentration of 5-HT in the prefrontal cortex was also statistically reduced by VPA treatment（t=2.145，P=0.049）. Conclusion：The both rat models with LPS or VPA treatment lead to autism-like behavors，whereas there are different induce molecular mechanisms. Therefore，we should make reasonable choice on the basis of research objectives in scientific study.