The photosynthetic fluorescence parameters of algae are easy to measure and sensitive to external stress. It is an important indicator of water quality biological toxicity measurement. However, different toxicity test schemes would lead to great differences in the toxicity data, and the stability and comparability were uncertain. Since fluorescence dynamics technology for measuring photosynthetic fluorescence parameters essentially uses optical detector to monitor the change of chlorophyll fluorescence signal in microalgae. In this paper, the concentration of chlorophyll was used as the measurement of biomass, and Chlorella pyrenoidosa was used as the test object to analyze the relationship between photosynthetic fluorescence parameters as the end point of toxicity test and the initial biomass of toxicity test algae samples under short-term exposure (1 h and 3 h) of diuron. The results showed that: 1) When the initial biomass changes, the correlation between photosynthetic fluorescence parameters and biomass would directly affect the stability of the toxicity test results. The photosynthetic fluorescence parameters could be divided into two categories. The first category was F_v/F_m, Yield, α, rP, σ_{PS II} and τ_es, whose values were not related to biomass changes and only represented the photosystem information of Chlorella pyrenoidosa. The toxicity test results obtained by these parameters were not affected by the change of initial biomass. When the initial chlorophyll concentration changed in the range of 20~1 000 μg·L^-1, 10 μg·L^-1 diuron stress for 1 h and 3 h, the average values of the corresponding relative standard deviations of the six parameter test results were 2.74% and 3.12% respectively, and the toxicity test results were stable. The second category of parameters was Ek, F₀, F_m, F_v and JVPⅡ. Their values were affected by biomass and contain biomass information. Among them, Ek was negatively correlated with chlorophyll concentration, and F₀, F_m, F_v and JVPⅡ were positively correlated with chlorophyll concentration. The toxicity test results of these parameter were obviously affected by biomass fluctuation and their stability became worse. When the chlorophyll concentration changed in the range of 20~1 000 μg·L^-1, the average values of relative standard deviation of the five parameter test results under 10 μg·L^-1 diuron stress for 1 h and 3 h were 14.66% and 17.27% respectively. 2) The Logistic model was selected for dose-effect analysis, and the photosynthetic fluorescence parameters F_v/F_m, Yield, α, rP, σ_PSⅡ, τ_es, F₀, F_m and F_v could establish a good dose-effect relationship with diuron. According to EC₅₀, EC₂₀ and correlation coefficient R², the optimal range of initial chlorophyll concentration of algae when photosynthetic fluorescence parameters were used as the end point of toxicity test were given. For parameters F_v/F_m, Yield, α, rP, σ_PSⅡ, and τ_es, due to population dependence, too low biomass would affect the photosynthetic physiological state of algae species, the recommended initial chlorophyll concentration range was 10~2 000 μg·L^-1; for parameters F₀, Fm and Fv, more sensitive toxicity test results (lower EC₂₀) could be obtained due to higher initial biomass, the recommended range was 200~1 000 μg·L^-1. This result provides an important basis for the establishment of a rapid detection method for water quality biological toxicity based on algal photosynthetic fluorescence parameters, and will be helpful for ecological risk assessment of the aquatic environment.