Steel trusses are widely used in civil and industrial structures thanks to their versatility and ability to accommodate various structural configurations. These structures exhibit complex behaviors, showing nonlinear inelastic responses affected by factors such as geometric and material nonlinearities, member interactions, and buckling. Traditional design techniques often fall short, leading to a rise in advanced analysis methods that, while more effective, require significant computational resources for complex tasks like seismic load analysis and structural optimization. To address this challenge, machine learning (ML) algorithms have been employed to predict steel truss behavior, improving decision-making efficiency and reducing error likelihood. This study evaluates eight ML algorithms, including three linear regressions and five tree-based ensemble algorithms, using a numerical example of a planar 113-bar truss bridge with 43 sectional design variables. Performance is assessed using four metrics: mean absolute error (MAE), mean squared error (MSE), coefficient of determination (R2), and computation time for each method.