Abstract: With n-octadecane, methyl methacrylate and chitosan as the core material, inner shell and outside shell respectively, the double-shell phase change and humidity storage micro-capsules were made by in-situ polymerization and emulsion interfacial polymerization method. Combining the uniform design and radial basis function neural network, the effect of parameters (i.e. emulsifier amount, co-emulsifier amount, initiator amount, phacoemulsification time and mixing speed) on particle size distribution of double-shell phase change and humidity storage micro-capsules were investigated. In addition, the optimal uniform particle size distribution of double-shell phase change and humidity storage micro-capsules were tested and characterized. The results show that radial basis function(RBF) neural network has the best approximation effect when "Spread" is 0.55. To obtain optimal uniform particle size distribution for double-shell phase change and humidity storage micro-capsules, the preparation parameters are identified as follows:emulsifier amount to be 3.58wt%, co-emulsifier amount to be 0.53wt%, initiator amount to be 1.86wt%, phacoemulsification time to be 11.35 min and mixing speed to be 632 r·min^-1. On the base of the above preparation parameters, d₁₀ is 3 352.0 nm, d₅₀ is 4 474.9 nm, d₉₀ is 6 108.4 nm, and the measured value of d₉₀-d₁₀ is 2 756.4 nm. Therefore, the measured value agrees well with the predictive value (relative error is 2.69%). In the case of relative humidity between 35% and 65%, the equilibrium moisture content is obtained to be between 0.0547 g·g^-1 and 0.1259 g·g^-1, phase transition temperature is 28.87℃ and phase change enthalpy is 78.45 J·g^-1.