Context. Propagating disturbances of the EUV emission intensity are commonly observed over a variety of coronal structures. Parameters of these disturbances, particularly the observed apparent (image-plane projected) propagation speed, are important tools for MHD coronal seismology. Aims. We design and test tools to reliably measure the apparent phase speed of propagating disturbances in imaging data sets. Methods. We designed cross-fitting technique (CFT), 2D coupled fitting (DCF) and best similarity match (BSM) to measure the apparent phase speed of propagating EUV disturbances in the running differences of time-distance plots (R) and background-removed and normalised time-distance plots (D).Results. The methods were applied to the analysis of quasi-periodic EUV disturbances propagating at a coronal fan-structure of active region NOAA11330 on 27 Oct. 2011, observed with the Atmospheric Imaging Assembly (AIA) on SDO in the 171 Å bandpass. The noise propagation in the AIA image processing was estimated, resulting in the preliminary estimation of the uncertainties in the AIA image flux. This information was used in measuring the apparent phase speed of the propagating disturbances with the CFT, DCF and BSM methods, which gave consistent results. The average projected speed is measured at 47.6 ± 0.6 km s −1 and 49.0 ± 0.7 km s −1 for R and D, with the corresponding periods at 179.7 ± 0.2 s and 179.7 ± 0.3 s, respectively. We analysed the effects of the lag time and the detrending time in the running difference processing and the background-removed plot, on the measurement of the speed, and found that they are fairly weak. Conclusions. The CFT, DCF and BSM methods are found to be reliable techniques for measuring the apparent (projected) phase speed. The samples of larger effective spatial length are more suitable for these methods. Time-distance plots with background removal and normalisation allow for more robust measurements, with little effect of the choice of the detrending time. Cross-fitting technique provides reliable measurements on good samples (e.g. samples with large effective detection length and recurring features). 2D coupled-fitting is found to be sensitive to the initial guess for parameters of the 2D fitting function. Thus DCF is only optimised in measuring one of the parameters (the phase speed in our application), while the period is poorly measured. Best similarity measure is robust for all types of samples and very tolerant to image pre-processing and regularisation (smoothing).