The IIoT also referred to as the Industrial internet of things helps in the transformation of daily life through improvement in efficiency and rise in productivity. But this is marred by some weaknesses or threats that are prone to some forms of malicious attacks. In regards to this, the work proposes ML models for the detection of such attacks. Our approach involves a comprehensive data preprocessing pipeline: For data pre-processing, it includes missing value imputation with mode and median, data transformations by creating columns like peak hour for attack from timestamp, frequency count for IPs and ports, removing some of the redundant columns, data encoding with One-Hot Encoder for nominal columns, and Label Encoding or map functions for ordinal columns., outlier analysis and treatment with IQR method, data scaling with Standard Scaler, and data selection with PCA and keeping the top 45 components. Following this, our proposed approach achieved outstanding results: Specificity = 0. 990, positive predictive value = 0. 997, sensitivity = 0. 983 and F1 measure = 0. 990. This model shows good capability in the identification of the attacks, thus protecting the IIoT system from such attacks.