Steel structures exposed to offshore atmospheric environment for a long time inevitably suffer from corrosion damage. Safety assessment of corroded steel structures largely depends on the quantification of corroded surface features as the irregular corrosion characteristics are the main factors causing decline in steel mechanical properties. To investigate the structural steel corrosion characteristics in offshore atmospheric environment, accelerated corrosion tests were conducted on 16 pieces of Q235B steel plates by periodic spraying to simulate the offshore atmospheric environment. Moreover, the surface morphologies and characteristic parameters were measured and analyzed using a ST400 3D Noncontact Profilometer and a self-written algorithm. The distribution characteristics such as corrosion depth, pit depth, and aspect ratio were elucidated, and the changing laws of statistical parameters such as mean value, standard deviation, and pitting shapes were revealed. The results indicated that in the simulated offshore atmospheric environment, the structural steel corrosion process generally goes through three stages: scab, swell, and spall. The scab and swell stages are dominated by pitting corrosion, whereas, the spall stage shows the general corrosion characteristics. Moreover, the corrosion depth of structural steel in the simulated offshore atmospheric environment conforms to the normal distribution, whereas, the pit depth and aspect ratio conform to the log-normal distribution. As the degree of corrosion increases, the mean value and standard deviation of the corrosion depth, peak value of the power spectral density of the corrosion depth, and logarithmic mean value of the pit depth also gradually increase, whereas, the logarithmic mean value of the pit aspect ratio decreases. Meanwhile, at different ages, the cone pits have the highest proportion, and the pit shape gradually changes from a cylinder or a hemisphere to a cone. Finally, based on the results of the statistical analysis of the corrosion depth and pit parameters, the stochastic field model of corrosion depth and random distribution model of corrosion pits were constructed, which achieved the accurate characterization and reproduction of the surface morphology of the corroded steel in a simulated offshore atmospheric environment. The research results would lay the foundation for the establishment of an accurate stochastic model and structural reliability analysis in the natural offshore atmospheric environment.