Nanoparticles of molecularly imprinted polymers (MIPs) were prepared by precipitation polymerization method. Glucose was used as a template molecule. The impact of different process parameters on the preparation of nanoparticles was investigated in order to reach the maximum binding capacity of MIPs. Experimental data based on uniform design were analyzed using artificial neural network to find the optimal condition. The results showed that the binding ability of nanoparticles of MIPs prepared under optimum condition was much higher than that of the corresponding non-imprinted nanoparticles (NIPs). The findings also demonstrated high glucose selectivity of imprinted nanoparticles. The results exhibited that the particle size for MIP nanoparticles was about 557.6 nm, and the Brunauer-Emmett-Teller analysis also confirmed that the particle pores were mesopores and macropores around 40 nm and possessed higher volume, surface area, and uniform size compared to the corresponding NIPs.