The sensitizing potential of chemicals is usually identified and characterized using in vivo methods such as the murine local lymph node assay (LLNA). Due to regulatory constraints and ethical concerns, alternatives to animal testing are needed to predict the skin sensitization potential of chemicals. For this purpose, an integrated evaluation system employing multiple in vitro and in silico parameters that reflect different aspects of the sensitization process seems promising. We previously reported that LLNA thresholds could be well predicted by using an artificial neural network (ANN) model, designated iSENS ver. 2 (integrating in vitro sensitization tests version 2), to analyze data obtained from in vitro tests focused on different aspects of skin sensitization. Here, we examined whether LLNA thresholds could be predicted by ANN using in silico-calculated descriptors of the three-dimensional structures of chemicals. We obtained a good correlation between predicted LLNA thresholds and reported values. Furthermore, combining the results of the in vitro (iSENS ver. 2) and in silico models reduced the number of chemicals for which the potency category was under-estimated. In conclusion, the ANN model using in silico parameters was shown to be have useful predictive performance. Further, our results indicate that the combination of this model with a predictive model using in vitro data represents a promising approach for integrated risk assessment of skin sensitization potential of chemicals.