We focus on empirically measure the p-factor of a homogeneous sample of 29 LMC and 10 SMC Cepheids for which an accurate average LMC/SMC distance were estimated from eclipsing binary systems. We used the SPIPS algorithm, which is an implementation of the BW method. As opposed to other conventional use, SPIPS combines all observables, i.e. radial velocities, multi-band photometry and interferometry into a consistent physical modeling to estimate the parameters of the stars. The large number and their redundancy insure its robustness and improves the statistical precision. We successfully estimated the p-factor of several MC Cepheids. Combined with our previous Galactic results, we find the following P-p relation: -0.08(log P-1.18)+1.24. We find no evidence of a metallicity dependent p-factor. We also derive a new calibration of the P-R relation, logR=0.684(log P-0.517)+1.489, with an intrinsic dispersion of 0.020. We detect an IR excess for all stars at 3.6 and 4.5um, which might be the signature of circumstellar dust. We measure a mean offset of $Delta m_{3.6}=0.057$mag and $Delta m_{4.5}=0.065$mag. We provide a new P-p relation based on a multi-wavelengths fit, and can be used for the distance scale calibration from the BW method. The dispersion is due to the MCs width we took into account because individual Cepheids distances are unknown. The new P-R relation has a small intrinsic dispersion, i.e. 4.5% in radius. Such precision will allow us to accurately apply the BW method to nearby galaxies. Finally, the IR excesses we detect raise again the issue on using mid-IR wavelengths to derive P-L relation and calibrate the $H_0$. These IR excesses might be the signature of circumstellar dust, and are never taken into account when applying the BW method at those wavelengths. Our measured offsets may give an average bias of 2.8% on the distances derived through mid-IR P-L relations.