Intracellular transport of large cargoes, such as organelles, vesicles or large proteins, is a complex dynamical process that involves the interplay of ATP-consuming molecular motors, cytoskeleton filaments and the viscoelastic cytoplasm. The displacements of particles or probes in the cell cytoplasm as a function of time are characterized by different (anomalous) diffusion regimes. We investigate here the motion of pigment organelles (melanosomes) driven by myosin-V motors in Xenopus laevis melanocytes using a high spatio-temporal resolution tracking technique. By analyzing the mean square displacement (MSD) of the obtained trajectories as a function of the time lag, we show that the melanosomes display a transition between subdiffusive to superdiffusive behavior. A stochastic theoretical model is introduced to generalize the interpretation of our data. Starting from a generalized Langevin equation that explicitly considers the collective action of the molecular motors we derive an analytical expression for the MSD as a function of the time lag, which also takes into account the experimental noise. By fitting our model to the experimental data we were able to discriminate the exponents that characterize the passive and active contributions to melanosome dynamics. The model also estimates the "global" motor forces correctly. In this sense, our model gives a quantitative description of active transport in living cells with a reduced number of parameters.