The similarities between the Fe-based superconductors and cuprate superconductors imply a possible unified picture of high temperature superconductivity. However, various chemical doping effects in Fe-based superconductors can lead to qualitatively similar phase diagrams that show diverse and complicated details, which pose great challenges of establishing a unified picture. Studying how chemical doping affects the electronic structure and superconductivity, and finding the real universal control parameter for superconductivity, are very important for establishing a unified picture and revealing the mechanism of high temperature superconductivity. In this article, we review a series of angle resolved photoemission studies on the chemical doping effect in Fe-based superconductors, involving both type I Fe-based superconductors with both electron and hole Fermi pockets, and type Ⅱ Fe-based superconductors with only electron Fermi pockets, and involving chemical doping of hetero-valent doping, isovalent doping, and chemical doping at different sites in unit cell. Comprehensive studies and analysis are conducted from various aspects of doping effects, including Fermi surfaces, impurity scattering, and electron correlation, and their roles in evolving the superconductivity. Electron correlation is found to be a universal electronic parameter behind the diverse phase diagrams of Fe-based superconductors, which naturally explains the qualitatively similar phase diagrams of various Fe-base superconductors despite of doping them in different ways. The electron correlation in Fe-based superconductors is closely related to both the carrier type of dopant and the lattice structure parameters, such as bond length. The different impurity scattering effects and different structures may affect the optimal Tc and thus leading to the diversity and complexity in the phase diagram. Fermi surface topology and its evolution with doping may play a secondary role in determining Tc. In order to enhance the Tc, one needs to optimize a moderate electronic correlation while minimizing the impurity scattering in the Fe-anion layer. Our results explain many puzzles and controversies and provide a new view for understanding the phase diagrams, resistivity behaviors, superconducting properties, etc. Our findings also strongly challenge the weak coupling theories based on the Fermi surface nesting, but favors the strong-coupling pairing scenario, where the competition between the electron kinetic energy and the local correlation interactions is a driving parameter of superconducting phase diagram. Like the t-J model of cuprates, in the picture of local antiferromagnetic exchange pairing, superconductivity appears in Fe-based superconductor when the electron correlation strength is at a moderate level. If the correlation is too weak, the system cannot exhibit superconductivity and remains metallic at low temperature. If the correlation is too strong, magnetic order appears in type I Fe-based superconductor, while type Ⅱ Fe-based superconductor shows a bandwidth-control correlated insulating state. The control parameter of the phase diagram is carrier doping for cuprates, but electron correlation strength for Fe-based superconductors. Our experimental results give a unified understanding of iron-based superconductors as a bandwidth-controlled system.