It was well recognized that mRNA lipid nanoparticles (LNPs) have significant stability challenges, due to the likelihood of mRNA degradation, nanoparticle aggregation, mRNA leakage from nanoparticles, hydrolysis, oxidation and the structure reformation etc. These make mRNA LNPs sensitive to the environment during storage and when exposed to body fluids. Stability evaluations are typically based on nanoparticle size, size distribution, zeta potential, and mRNA encapsulation. However, we found that these parameters had no significant change when subjected to various storage conditions, while the transfection efficacy of mRNA was largely different. We compared the mRNA LNPs stored as liquid, frozen and lyophilized form with different lyoprotectants, under various storage conditions. We found the use of cell lines and reporter mRNA need to be carefully selected. When using mRNAs encoding firefly luciferase (FLuc mRNA), the transfection differs between cell lines and showed non-linear correlation between mRNA concentration and transfection efficiency in cultured cells. However, mRNA encoding GFP is much more stable with proportional dose response. On the other hand, the amount of mRNA internalized in cells was found not directly corrected with transfection, indicating the morphology and packing changes of LNPs may affect the release of mRNA to cytoplasm. Further structure analysis through small angle x-ray scattering and cryoEM confirmed the changes in particle morphology and lipid packing after lyophilization. The talk will use the recent data to illustrate that particle characterization based on physicochemical properties may not reflect the function of the mRNA LNPs. The structure and function relation is critical for understanding the mRNA LNP stability.