Past investigations have shown the sensitivity of masonry structures to record-to-record variability (aleatory uncertainty). In this regard, the selection of seismic input for performance-based assessment might be an issue in regions with a lack of recorded accelerograms characteristic of large-magnitude events, which can be solved by utilising simulated records. This paper examines the consistency of fragility functions of masonry archetypes using real and simulated signals as input for non-linear incremental dynamic analyses (IDAs). Structural models are developed based on representative archetypes and typologies of the residential building stock in Portugal. These buildings are modelled in OpenSees to consider In-Plane (IP) and Out-of-Plane (OOP) effects simultaneously while improving computational efficiency. Ground motion simulations are carried out using a stochastic finite-fault simulation approach that incorporates the dynamic corner frequency concept. Following this, two sets of records are independently selected, consisting of both real and simulated accelerograms. The real records are chosen to meet seismological compatibility criteria with the simulated data. Subsequently, IDAs are performed, and fragility curves are developed. The fragility obtained after real and simulated records are contrasted in terms of the moments of the distribution (median and dispersion). Finally, some practical recommendations are provided for the utilisation of simulated ground motions for performance assessment of masonry structures.