Abstarct: Bits are one of the most important parts of drilling, due to the type of application and high costs. Wear of these tools are an important parameter for estimating the efficiency of drilling equipment in mining projects. Examining the parameters affecting the tricone rotary bits wear can prevent wasting time and additional costs of the drilling process. Therefore, in this study, statistical methods and deep learning have been used to predict the tricone rotary bits wear. In the first part, the tricone rotary bits wear has been calculated through weight loss, dimensional loss, and volumetric loss, and by determining the effective parameters baxsed on the results of univariate analysis and Principal Component Analysis (PCA), the relationship between these factors and the tricone rotary bits wear has been determined through statistical methods. In this study, only 21 models, including 6 linear models and 15 nonlinear models, have been approved to predict the rate of tricone rotary bits wear while drilling. Also, due to recent advances in the use of deep learning in the field of computer vision and the proper performance of these methods, in the second part of the research, deep learning has been used to predict the tricone rotary bits wear. Due to the low number of images taken from drilling bits and the imbalance in the distribution of images of transfer learning baxsed on the extracted features, training on the large ImageNet dataset has been used. Then the 16-laxyer VGG architecture is selected as the appropriate architecture according to the proper performance in the validation and test set. Then, due to the dimensions and high number of mappings of the extracted features, according to PCA, 35 features were used to predict the test set. Finally, to predict the wear rate of tricone rotary bits from regression-baxsed machine learning models were usedIt should be noted that the performance of the models has been examined baxsed on four evaluation criteria. Then, using the three strategies of prioritizing the mean ranks, Borda and Copeland’s method, finally using the arithmetic mean, the prioritization of the models has been done by aggregating the ranking of the above methods. The rank aggregation method shows that the best model for predicting the tricone rotary bits wear is the nonlinear regression model in which the independent variables include Weight On Bit (WOB), Uniaxial Compressive Strength (UCS) and Geological Strength Index (GSI). Also, baxsed on the performance of the results obtained from the machine learning prediction models ranking, Gaussian Process Regression (GPR) has been selected as the best model for predicting the tricone rotary bits wear.