Sediment is an important factor involved in sustainable development of water resources systems. Suspended sediment is a major portion of total load transported by streams. Quantification of annual maximum suspended sediment load (SSLm) and its quantiles (SSLm*) is the basic requirement for planning, design and management of hydraulic structures and water resources-related projects. Less sediment sampling is very common in developing and remote regions and it is the main constraint for extreme event study. This research attempts to estimate site-specific SSLm* based on discontinuous time series with low sampling frequency and develop regional empirical approaches for predicting SSLm* of ungauged tributaries in the Lower Mekong Basin. Firstly, SSLm were distinguished from the observed records. Due to data insufficiency for fitting probability distribution, two methods were then applied to estimate SSLm for each station, but only the one using regional linear regression was valid. The site-specific SSLm* could be estimated with an accuracy of an average mean absolute percentage error of 35%. Finally, empirical equations associated with catchment area with slope greater than 15% (A15) were established and can be used to estimate SSLm* until 100-year return period with A15 ranging between 120 and 6,540 km2.