The regional data assimilation system at the Japan Meteorological Agency uses a variational data assimilation system based on the non-hydrostatic model ASUCA (named ASUCA-Var). This paper reviews the configurations and current status of ASUCA-Var. To consider the consistency of analysis and prognostic variables, the control variables of ASUCA-Var include soil variables and basic atmospheric variables. The background-errors based on the control variables are calculated every 3 h for land and sea grid points to better reflect the representative error covariance structure, considering daily variations and differences in structure on land and sea. Although the cost function is designed to be a perfect quadratic form, the basic field update method in the optimization process allows the nonlinearity of the observation operator and numerical weather prediction model to be incorporated into the solution of optimization problem in the incremental four-dimensional variational (4D-Var) method. The outer/inner models used in the incremental 4D-Var method are based on ASUCA, with suitable configurations according to each resolution and applied linearization. Observation operators are implemented for various kinds of observations used, with unified interfaces encapsulating external simulators. Variational quality control and variational bias correction are also introduced for advanced observation handling within the variational system. Parallelization is introduced to enhance computational efficiency, including adjoint calculations. To assess the impact of assimilated observations, degrees of freedom for signal are also available. Additionally, as a system for operational use, ASUCA-Var is designed for sustainable development. The meso-scale analysis and local analysis workflows are presented as operational implementations of ASUCA-Var. ASUCA-Var enhances forecasting in a wide range of validation indices. The major future improvements of ASUCA-Var include the introduction of the flow-dependent background-error and extension of the control variable to hydrometeors, which are expected to enhance the prediction accuracy of the operational regional model.