The fast release rate and the undesirable covalent binding are two major problems often encountered in formulating in situ chemically cross-linked hydrogel as protein release depot, particularly when prolonged release over months is desirable. In this study, we applied the De Gennes blob theory to analyze and tackle these two problems using a vinylsulfone-thiol (VS-SH) reaction based in situ hydrogel system. We showed that the simple scaling relation xi b approximate to R-g(c/c*)-v/(3v-1) is applicable to the in situ hydrogel and the mesh size estimated from the precursor polymer parameters is a reasonable match to experimental results. On the other hand, as predicted by the theory and confirmed by experiments, the drug diffusion within hydrogel depends mainly on polymer concentration but not the degree of modification (DM). The covalent binding was found to be caused by the mismatch of location between the reactive groups and the entanglement points. The mismatch and, thus, the protein binding were minimized by increasing the DM and concentration of the SH polymer relative to the VS polymer, as predicted by theory. Using these principles, an in situ hydrogel system for the controlled release of an antiangiogenic antibody therapeutics bevacizumab for 3 months was developed.