The friction force cannot be predicted accurately by using the classical friction model without considering hydraulic fluid pressure effects. In order to solve this problem, the modified steady-state friction model (P-Stribeck) and dynamic friction model (P-GMS) based on Stribeck and generalized Maxwell slip model (GMS) were proposed by introducing pressure influence coefficient dynamic friction time constant. To validate the effectiveness of those proposed friction models, a hydraulic cylinder test platform was developed to study the friction characteristics of hydraulic cylinders. Therefore, the friction behaviors of hydraulic cylinders for different seal types, different cylinder bores, different loads, different velocity and frequencies were investigated. The intelligent genetic algorithm was adopted to identify the friction parameters of the proposed steady friction model and dynamic friction model. The identification and validation procedures of the developed friction models were conducted with the data of inlet and outlet pressure, displacement/velocity, friction force and other data collected from the hydraulic cylinder test bench. The experimental results were compared with that of predicted by the classic friction models and proposed friction models. Also, an error analysis procedure of the proposed friction models was conducted under different operation conditions. The results showed that the predicting accuracy of steady state friction force of the P-Stribeck model was better than that of the Stribeck model;in addition, the precision of friction estimation of the proposed P-GMS model was higher than that of the GMS model.