Diversity patterns play an important role in identifying the mechanisms of biodiversity maintenance. The soil of the Ebinur Lake Wetland Nature Reserve located in the western margin of the Gurbantonggut Desert, in the Xinjiang Uygur Autonomous Region of China, is short in water and rich in salinity. These regional conditions result in plants that have developed high sensitivity and vulnerability characteristics. Thus, the exploration of desert plant diversity patterns and response mechanisms along various soil water and salinity gradients can integrate community-wide information to reveal the mechanisms of plant adaptations in an arid region and provide scientific strategies for reserve management and biodiversity conservation. However, plant diversity patterns and responses to varying gradients of soil water and salinity in the Ebinur Lake remain unclear. In the present study, three transects were first established that were perpendicular to the north bank of the Aqikesu River (one of three rivers in the Ebinur Lake Wetland Nature Reserve). In each of the three transects, 10—12 plots (total of 32 plots) were set. The geographic data, abundance, richness, and soil volume water content (SVWC) were surveyed and recorded in each plot. The soil samples were then collected and taken to the Key Laboratory of Oasis Ecology of Xinjiang University. All soil samples were air dried, and soil pH (pH), electrical conductivity (EC), soil organic carbon (C), soil total nitrogen (N), soil total phosphorus (P), and soil total sulfur (S) of all samples were measured in the laboratory. Based on these values, abundance models, multiple comparisons, ordination analysis, and the regression method were used to explore plant abundance distributions, plant diversity patterns, and relationships with the soil environment across soil water and salinity gradients. The results showed that: (1) SVWC and EC in high (SW1) and low (SW2) soil water and salinity plots were 16.65% and 12.02 mS/cm; and 2.63% and 1.91 mS/cm, respectively. Species composition of herb and some shrub communities were mainly influenced by the variation in soil water and salinity. (2) The plant distribution pattern showed no significant relationship with soil environmental factors in SW1 plots; whereas in SW2 plots, significant variations were noted with SVWC, EC, and soil nutrient content, which presented a pattern of regional distribution. (3) Plant abundance patterns were fitted using the lognormal (LN) and Zipf models for SW1 and SW2, respectively. Variability in plant abundance patterns indicated the responses of community composition and structure to soil water and salinity in arid desert conditions. (4) Plant diversity in SW1 was significantly higher than that in SW2. To some extent, this diversity was significantly affected by soil pH, SVWC, and S. Overall, the relationships between distribution and diversity patterns, and the soil environment showed some response patterns to soil water and salinity gradients. In addition, the responses of a plant community to the soil environment across soil water and salinity gradients can provide reference points for the control of vegetation restoration and soil salinization under arid desert conditions.