We compare the results of two different approaches in Local Earthquake Tomography (LET) as applied to the Alban Hills volcano. The first approach is called the graded inversion approach in which we progressively focused towards the center of the volcano while decreasing the model grid spacing in subsequent inversions, optimizing gradually the velocity model. The second approach is called the normal approach in which the seismic arrival times were directly inverted using a ID velocity model optimally representing the background structure. The target region, i.e. Alban Hills volcano (or Vulcano Laziale) is a Quaternary age volcano lying about 20 kin southeast of Rome, Italy. A microseismic network temporarily installed on and around the volcano during a quake swarm gathered the arrival time data that we used. In the normal approach, we used three different grid spacings (i.e. 2, 1.5, and I km) that provided detailed images of the volcano. The resolution analysis carefully performed on the model parameters allowed the determination of a more reliable final model that represented the best results for the velocity structure beneath the volcano. The inversion results attained with the graded approach were not satisfactory, as some small-scale heterogeneities were not properly detected as done by the inversion with the normal approach. The normal approach in fact revealed a horn shaped structure with a high velocity located beneath the volcano and a low velocity anomaly dominated the depths around 1-4 kin in western side of the volcano. In general, the graded approach relative to the normal approach failed to detect the fine details of the volcano.