The present work evaluates the influence of the installation in ultrasonic flow meters, using numerical simulation. Ultrasonic meter characteristics and issues about the flow modeling configuration that may influence the numerical results were presented. Procedures used to draw the lines and to calculate the meter mean velocity and the measurement error were described. The effect of one curve, two curves at the same plane and two curves at perpendicular planes at 1, 2, 3, 4, 5 and 8-path meters with different arrangements were shown. Also, the effect of one step at the results of ultrasonic meters was analyzed by the simulation of 1 to 5-paths meters and by the calibration of a 3 path meters. Factors calculated by the ratio between the velocity indicated by the meter in a certain position and the velocity that would be indicated by the meter in a fully developed flow. Just after one curve, the factors for 1-path meters varied from 0.40 to 1.28; for 3-path meters, from 0.5 to 1.42; and for 4 and 5-path meters, the differences were higher than 10%. At 20D downstream of one curve, 2 and 3-path meters indicated factors equal to 0.9 and 1.08; and 4 and 5-path meters, factors equal to 0.99 to 1.04. The diagnostic parameters analyzed were not efficient and their use is not recommended. By the simulation, a convergent step of -4% generates differences about 0.3% for 3-path meters and for a divergent step -f 4%, the differences are higher than 0.5%. The differences between the calibration results with and without steps were from 0.18% to 0.3%. The numerical simulation is a useful tool in the ultrasonic meter analyses and showed the sensitivity of these meters to velocity profile variations.