A method called LipoP was developed to predict lipoprotein signal peptides in Gram-negative bacteria using a hidden Markov model (HMM) and a neural network. The HMM outperformed existing methods, correctly predicting 96.8% of lipoproteins with only 0.3% false positives. It also accurately identified 92.9% of lipoproteins in Gram-positive bacteria. The method was tested on 12 Gram-negative and one Gram-positive genome, showing good agreement with experimental data. The HMM was able to distinguish between lipoproteins, SPaseI-cleaved proteins, cytoplasmic proteins, and transmembrane proteins. It also predicted cleavage sites in both SPaseI- and SPaseII-cleaved signal peptides. The HMM was compared with a neural network, which gave similar results. The HMM was found to be more accurate than existing methods, with fewer false positives. The method was used to predict lipoproteins in 12 Gram-negative bacteria, with 94.6% of experimentally verified lipoproteins correctly predicted. The HMM was also able to predict Gram-positive lipoproteins, showing its versatility. The LipoP server is available online for genome-wide predictions. The study highlights the effectiveness of HMMs in predicting lipoprotein signal peptides and their cleavage sites, providing a valuable tool for bacterial protein analysis.A method called LipoP was developed to predict lipoprotein signal peptides in Gram-negative bacteria using a hidden Markov model (HMM) and a neural network. The HMM outperformed existing methods, correctly predicting 96.8% of lipoproteins with only 0.3% false positives. It also accurately identified 92.9% of lipoproteins in Gram-positive bacteria. The method was tested on 12 Gram-negative and one Gram-positive genome, showing good agreement with experimental data. The HMM was able to distinguish between lipoproteins, SPaseI-cleaved proteins, cytoplasmic proteins, and transmembrane proteins. It also predicted cleavage sites in both SPaseI- and SPaseII-cleaved signal peptides. The HMM was compared with a neural network, which gave similar results. The HMM was found to be more accurate than existing methods, with fewer false positives. The method was used to predict lipoproteins in 12 Gram-negative bacteria, with 94.6% of experimentally verified lipoproteins correctly predicted. The HMM was also able to predict Gram-positive lipoproteins, showing its versatility. The LipoP server is available online for genome-wide predictions. The study highlights the effectiveness of HMMs in predicting lipoprotein signal peptides and their cleavage sites, providing a valuable tool for bacterial protein analysis.