Deterioration models can be used to forecast the evolution of the condition of the sewer network under different investment strategies. Models are calibrated using condition scores obtained from sewer visual inspection. Many studies highlighted the uncertainties in the procedure of sewer condition assessment, mainly due to the subjectivity of the coding operator. However, the influence of this uncertainty on the outcomes of deterioration modelling remains unknown. This paper analyses the influence of sewer condition uncertainties on the prediction of deterioration models. An optimisation methodology has been applied to quantify uncertainties in sewer condition assessment from the analysis of a set of repeated inspections. Then, a method is proposed to propagate uncertainties in the survival curves of a deterioration model using the inspection dataset of the city of Berlin, Germany. Results indicate that old pipes in bad condition are more prone to False Negative than False Positive (higher probability to miss defects). As a result, the propagation of uncertainties leads to more pessimistic survival curves with a confidence interval of ±12% at 100 years. The analysis shows that the required replacement rate to maintain a constant network condition is underestimated if uncertainties are not considered.