Phosphorus (P) loss from agricultural fields is of environmental concern because of its potential impact on water quality in streams and lakes. The Mississippi Delta has long been known for its fish productivity and recreational value, but high levels of P in fresh water can lead to algal blooms that have many detrimental effects on natural ecosystems. Algal blooms interfere with recreational and aesthetic water use. However, few studies have evaluated P losses from agricultural watersheds in the Mississippi Delta. To better understand the processes influencing P loss, rainfall, surface runoff, sediment, ortho-P (orthophosphate, PO₄–P), and total P (TP) were measured (water years 1996–2000) for two subwatersheds (UL1 and UL2) of the Deep Hollow Lake Watershed and one subwatershed of the Beasley Lake Watershed (BL3) primarily in cotton production in the Mississippi Delta. Ortho-P concentrations ranged from 0.01 to 1.0 mg/L with a mean of 0.17 mg/L at UL1 (17.0 ha), 0.36 mg/L at UL2 (11.2 ha) and 0.12 mg/L at BL3 (7.2 ha). The TP concentrations ranged from 0.14 to 7.9 mg/L with a mean of 0.96 mg/L at UL1, 1.1 mg/L at UL2 and 1.29 mg/L at BL3. Among the three sites, UL1 and UL2 received P application in October 1998, and BL3 received P applications in the spring of 1998 and 1999. At UL1, ortho-P concentrations were 0.36, 0.25 and 0.16 for the first, second and third rainfall events after P application, respectively; At UL2, ortho-P concentrations were 1.0, 0.66 and 0.65 for the first, second and third rainfall events after P application, respectively; and at BL3, ortho-P concentrations were 0.11, 0.22 and 0.09 for the first, second and third rainfall events after P application, respectively. P fertilizer application did influence P losses, but high P concentrations observed in surface runoff were not always a direct result of P fertilizer application or high rainfall. Application of P in the fall (UL1 and UL2) resulted in more ortho-P losses, likely because high rainfall often occurred in the winter months soon after application. The mean ortho-P concentrations were higher at UL1 and UL2 than those at BL3, although BL3 received more P application during the monitoring period, because P was applied in spring at BL3. However, tillage associated with planting and incorporating applied P in the spring (BL3) may have resulted in more TP loss in sediment, thus the mean TP concentration was the highest at BL3. Ortho-P loss was correlated with surface runoff; and TP loss was correlated with sediment loss. These results indicate that applying P fertilizer in the spring may be recommended to reduce potential ortho-P loss during the fallow winter season; in addition, conservation practices may reduce potential TP loss associated with soil loss.