A clear understanding of the basic forms of motion of charged particles in the magnetic and electric fields of the outer magnetosphere is an essential ingredient to the study of the earth's radiation belts and plasma. To achieve this, one must work with a reliable quantitative model of the magnetospheric field. The main purpose of any such model is to provide a mathematical description of the field that is reasonably accurate within a given spatial and temporal domain and that can be used to analyze in a quantitative way any process influenced or governed by this field in the corresponding domain. In this review, magnetic field models of the magnetosphere are critically examined and their fit to in‐situ field measurements is discussed. Special attention is paid to the analysis of typical long‐term time variations of the field configuration and the matching changes in model parameters. Implications for adiabatic particle motion, such as shell geometry, regions of pseudo‐trapping, and adiabatic effects of slow time variations of the field, are ana