In this paper, an attempt to analyse landslide hazard and vulnerability in the municipality of Pahuatlán, Puebla, Mexico, is presented. In order to estimate landslide hazard, the susceptibility,magnitude(area-velocity ratio) and landslide frequency of the area of interest were produced based on information derived from a geomorphological landslide inventory; the latter was generated by using very high resolution satellite stereo pairs along with information derived from other sources(Google Earth,aerial photographs and historical information).Estimations of landslide susceptibility were determined by combining four statistical techniques:(i) logistic regression,(ii) quadratic discriminant analysis,(iii) linear discriminant analysis, and(iv)neuronal networks. A Digital Elevation Model(DEM)of 10 m spatial resolution was used to extract the slope angle, aspect, curvature, elevation and relief.These factors, in addition to land cover, lithology anddistance to faults, were used as explanatory variables for the susceptibility models. Additionally, a Poisson model was used to estimate landslide temporal frequency, at the same time as landslide magnitude was obtained by using the relationship between landslide area and the velocity of movements. Then,due to the complexity of evaluating it, vulnerability of population was analysed by applying the Spatial Approach to Vulnerability Assessment(SAVE) model which considered levels of exposure, sensitivity and lack of resilience. Results were expressed on maps on which different spatial patterns of levels of landslide hazard and vulnerability were found for the inhabited areas. It is noteworthy that the lack of optimal methodologies to estimate and quantify vulnerability is more notorious than that of hazard assessments.Consequently, levels of uncertainty linked to landslide risk assessment remain a challenge to be addressed. In this paper, an attempt to analyze landslide hazard and vulnerability in the municipality of Pahuatlán, Puebla, Mexico, is presented. In order to estimate landslide hazard, the susceptibility, magnitude (area-velocity ratio) and landslide frequency of the area of ​​interest were produced based on information derived from a geomorphological landslide inventory; the latter was generated by using very high resolution satellite stereo pairs along with information derived from other sources (Google Earth, aerial photographs and historical information). Estimations of landslide susceptibility were determined by combining (ii) quadratic discriminant analysis, (iii) linear discriminant analysis, and (iv) neuronal networks. A Digital Elevation Model (DEM) of 10 m spatial resolution was used to extract the slope angle , aspect, curvature, elevation and relief.These factors, in addition to land cover, lithology and distance to faults, were used as explanat ory variables for the susceptibility models. Additionally, a Poisson model was used to estimate landslide temporal frequency, at the same time as landslide magnitude was was used by the relationship between landslide area and the velocity of movements. Then, due to the complexity of evaluating it, vulnerability of population was analysed by applying the Spatial Approach to Vulnerability Assessment (SAVE) model which considers levels of exposure, sensitivity and lack of resilience. Results were expressed on maps on which different spatial patterns of levels of landslide hazard and vulnerability were found for the inhabited areas. It is noteworthy that the lack of optimal methodologies to estimate and quantify vulnerability is more notorious than that of hazard assessments. Conclusion, levels of uncertainty linked to landslide risk assessment remain a challenge to be addressed.