In data assimilation for weather forecast, ensemble Kalman filter assumes linearity of the observation operator and Gaussianity of the probability distribution function (PDF) to explicitly solve the analysis. As a method avoiding errors based on these assumptions, we describe a four-dimensional ensemble-based variational method (4D-EnVAR) with observation localization. This formulation differs from that of the four-dimensional local ensemble transform Kalman filter (4D-LETKF) only in two points: (1) not assuming linearity of the observation operator and (2) calculating it globally. Using single-observation assimilation experiments and the observation system simulation experiments with a low-resolution atmospheric general circulation model, we demonstrate that 4D-EnVAR with observation localization has an advantage over 4D-LETKF because the observation operator is globally calculated in EnVAR.(Citation: Yokota, S., M. Kunii, K. Aonashi, and S. Origuchi, 2016: Comparison between four-dimensional LETKF and ensemble-based variational data assimilation with observation localization. SOLA, 12, 80−85,