森林生物量是森林生态系统的最基本数量特征 ,生物量数据是研究许多林业问题和生态问题的基础 ,因此 ,准确测定生物量十分重要。建立生物量模型是生物量估测的主要手段。以往所建模型 ,存在一个严重的缺陷 ,即各分量模型间不相容。如何解决相容性问题 ,一直是生物量估计领域所面临的一个难题。本文以长白落叶松为实例 ,提出了一种新方法———非线性联合估计法 ,并与比例平差法进行了对比。针对不同建模方法 ,设计了 5种估计方案 ,经过分析比较 ,确定了 1种方案为最优估计方案。该方案以树干生物量作为控制量 ,采取两级联合估计。模型构成如下第一级,W1=f2(x)+f5(x),W2=f2(x),W5=f5(x);;第二级,W3=f3(x),W4=f2(x)-f3(x);;第三级,W6=f6(x),W7=f5(x)-f6(x)。本文中模型选型采用了变量逐步筛选法,参数估计采用了加权最小二乘法,以消除异方差现象。同时,提出了5个指标用于模型评价,即参数变动系数C%、总相对误差RS%、平均相对误差EE%,平均相对误差绝对值RMA%和预估精度P% Forest biomass is the most basic quantitative characteristic of forest ecosystem. Biomass data is the basis for studying many forestry and ecological problems. Therefore, it is very important to measure biomass accurately. The establishment of biomass model is the main means of biomass estimation. The model built in the past, there is a serious flaw, that is, between the component model is not compatible. How to solve the compatibility problem has always been a challenge in the field of biomass estimation. In this paper, an example is given to the case of Changbai Larch, a new method - nonlinear joint estimation method is proposed, which is compared with the proportional adjustment method. According to different modeling methods, five kinds of estimation schemes are designed. After analysis and comparison, one scheme is determined as the optimal estimation scheme. The program uses the trunk biomass as the control volume and adopts a two-level joint estimation. The model consists of the first stage with W1 = f2 (x) + f5 (x), W2 = f2 (x) and W5 = f5 ) -f3 (x) ;; third level, W6 = f6 (x), W7 = f5 (x) -f6 (x). In this paper, the model selection using variable step-by-step screening method, the parameter estimation using weighted least squares method to eliminate heteroscedasticity. At the same time, five indexes are put forward for the model evaluation, namely the parameter variation coefficient C%, the total relative error RS%, the average relative error EE%, the average relative error absolute value RMA% and the prediction accuracy P%