In existing simulations of the Almen intensity test, arc height is indirectly obtained by an equivalent method including a representative cell, a few shots and equivalent loading. Most of these equivalent methods cannot consider the transverse deformation of the strip, the complex stress state of the plastic hardening layer and process parameters, resulting in deviation from the actual test. This paper introduces an improved and experimentally validated discrete element model (DEM)-finite element model (FEM) to predict the actual Almen intensity. The improvement of this model is mainly reflected in the large and real number of shots involved in the actual Almen intensity test, shot-shot interactions, and real-size solid finite element model of the Almen strip. A new method for calculating the shot stream is proposed based on the test and considering test process parameters such as the mass flowrate, nozzle movement speed and nozzle-workpiece distance. The shot stream impacting the strip with a fully restrained underside was first simulated in improved DEM-FEM to bring the forming energy. As a second step, an implicit solver of the Almen strip FEM calculates the spring-back to simulate strip removal from the holder. The results achieved by the present approach are compared with the results obtained by the experimental results and those in the literature. The results show that the arc height and Almen intensity obtained by the present approach match much better with the literature than the traditional method. Some new results obtained by the improved coupling DEM-FEM method are presented. The influences of the transverse deformation and surface plastic layer on the deformation of the Almen strip are discussed. This improved method provides an alternative characterization method for precision peen forming simulation.