Negative SU-8 photoresist processed at several levels of lower temperature than conventional approach was investigated, and its low-temperature processing has been characterized in terms of delamination. According to two phases of statistical design of experiment (DOE), initially statistically significant variables of SU-8 photoresist processing were selected, and the selected variables were further investigated to find their effects on delamination of SU-8 microstructure. A three leveled factorial designed experiments were performed followed by a 2(6-1) fractional factorial designed as a screening experiment. Characterizing low-temperature SU-8 photoresist process, 27 samples were fabricated and the degree of delamination was measured. In addition, nine additional samples were fabricated for the purpose of verification. Employing a neural network modeling technique, a process model is established, and response surfaces are generated to investigate the degree of delamination associated with three process parameters: post exposure bake (PEB) temperature, PEB time, and exposure energy. From the response surfaces generated, two significant parameters associated with delamination are identified, and their effects on delamination were analyzed. The higher the post exposure bake (PEB) temperature at a fixed PEB time and the more delamination occurred. In addition, the higher the dose of exposure energy, the lower the temperature at which the delamination begins and the larger the degree of delamination. The results identified acceptable ranges of the three process variables to avoid the delamination of SU-8 film, which in turn might lead to potential defects in MEMS device fabrication.