Metal oxides (MeOx) are an important group of engineered nanoparticles (NPs) as they are extensively used in cosmetics, textiles, coatings, pesticides, medicines as well as in wastewater treatment. These metal oxides can cause deleterious effects on human and environmental health which need to be considered before their production as well as release to the environment. Limited experimental data on the toxicity profile of MEOx NPs to diverse organisms of different trophic levels being available, in silico prediction tools appear to be an alternative non-animal method (NAM) for effective data gap bridging. As the number of data points available for experimental nanotoxicity data is in general very limited, the development of statistically robust and reliable quantitative structure-activity relationship (QSAR) models is challenging. On the other hand, read-across, a similarity-based technique, being originally a non-statistical approach has gained importance in efficient nanotoxicity predictions. We have recently developed a new quantitative read-across methodology with the incorporation of machine learning optimization of different hyperparameters for nano-read-across predictions of toxicity and property endpoints of newly synthesized NPs based on the similarity (Euclidean distance-based similarity, Gaussian kernel function similarity, Laplacian kernel function similarity) with structural analogs. A java based program (available from https://sites.google.com/jadavpuruniversity.in/dtc-lab-software/home) has also been developed based on the proposed algorithm which can effectively predict the toxicity of unknown NPs after providing the structural information of chemical analogs. The quality of predictions depends on the selection of the distance threshold, similarity threshold, and the number of most similar training compounds. The quality of quantitative predictions can be judged by using various correlation and error–based validation metrics. As the confidence of predictions for untested compounds is important information, we have addressed this issue by consideration of several criteria such as weighted standard deviation of the predicted values, coefficient of variation of the computed predictions, average similarity level of close training compounds for each query molecule, standard deviation and coefficient of variation of similarity levels, maximum similarity levels to positive and negative close training compounds, a perplexity measure indicating similarity to positive, negative or both classes of close training compounds, etc. It is also possible to club QSAR and read-across techniques to develop Read-across Structure-Activity Relationship (RASAR) models using the chemical similarity concepts of read-across (unsupervised step) and finally develop a supervised learning model (like QSAR). Nano-read-across and nano-RASAR appear to be promising approaches for risk assessment of nanomaterials complying with ethical considerations. Several case studies of such successful applications will be discussed.