The harvesting of Mozambique tropical hardwood species is considerable higher than the natural regrowth in the forest and the stock is decreasing drastically. Therefore, it is important to improve the material recovery when the wood is refined (i.e., in sawing and further refinement to products such as joineries, furniture etc.) to reduce the waste and to re-utilize efficiently the by-products to increase the added value. The wood processing industry is an important means to boost the industries in the rural areas and also to generate incomes for the local communities by creating jobs and business opportunities. The majority of the logging that can be used for sawmilling in Mozambique is exported as roundwood due to the inability of companies to meet the product standards set for export and to generate profit. The lack of capabilities of the local sawmills to generate profit, also foments the illegal logging because of the higher price of roundwood for export which contributes to increase the number of unlicensed individuals in harvesting. This threatens the law enforcement and thus the degradation of the local wood industry. An alternative to increase the profit and empower the local community could be to export more refined wood products such as sawn timber, parquet, and veneer instead of the roundwood. The objective of the work was to investigate alternative sawing strategies of tropical hardwood species that could increase the profitability of the Mozambique wood industry in general and at sawmill in particular. The subject was approached using a database of virtual logs and together with a sawing simulator. The thesis has two main focus areas: (1) creating the log database with the corresponding algorithms for sawing simulation, and (2) investigations of alternative sawing strategies. The first focus was to build the database of surface-scanned logs and develop the algorithm for the saw simulation. The results are a database of 15 logs models describing the logs outer shape in which 10 jambirre (Millettia stuhlmannii Taub.) and 5 umbila (Pterocarpus angolensis DC.), and the algorithm for the sawing simulation. The algorithm use “brute force” i.e., determines all volume yields of sawn timber from the combination of all settings of log-positioning parameters (offset, skew and rotation) and selects the maximum value of volume yield. From simulation, using three sawing patterns (cant-sawing, through-and-through sawing and square-sawing) combined with two positioning parameters (offset and rotation) it was found that the sawing pattern has great impact on volume yield and that the square-sawing gave higher yield followed by through-and-through sawing pattern. The second focus was on alternative sawing strategies; having in mind that the optimal volume yield is achieved by aid of computerized production systems and that these resources are not yet in use in Mozambique. Hence, the objective was to find the positioning parameters that can be set manually and improve the volume yield. The result have shown that the rotation is the most affecting parameter followed by offset and skew, and that the volume yield can decrease by between 7.7% and 12.5% from the optimal positioning when the logs are manually positioned with the knowledge about the optimal log position. In another study, using crook-up or horns-down positioning as alternative to the optimal positioning, the volume yield decreases by between 10% and 13% from the optimal positioning. By using bucked logs , the optimal volume yield increased by between 8% and 13% in relation to full lengths logs, and the volume yield of bucked logs when using crook up positioning decreases 2% in relation to optimal positioning of full length logs. It is concluded that there is an unexploited value potential in the wood chain which can be reached using alternative positioning and modern measurement techniques and that the grading of wood will facilitate and improve the sawing process.